{"id":62237,"date":"2022-08-12T18:07:22","date_gmt":"2022-08-12T18:07:22","guid":{"rendered":"https:\/\/loopbeatzz.info\/wp\/image-thresholding-segmentation-based-on-weighted-parzen-window-and-linear-programming-techniques-scientific-reports-nature-com\/"},"modified":"2022-08-12T18:07:22","modified_gmt":"2022-08-12T18:07:22","slug":"image-thresholding-segmentation-based-on-weighted-parzen-window-and-linear-programming-techniques-scientific-reports-nature-com","status":"publish","type":"post","link":"https:\/\/loopbeatzz.info\/wp\/image-thresholding-segmentation-based-on-weighted-parzen-window-and-linear-programming-techniques-scientific-reports-nature-com\/","title":{"rendered":"Image thresholding segmentation based on weighted Parzen-window and linear programming techniques | Scientific Reports &#8211; Nature.com"},"content":{"rendered":"<p>Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain             the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in             Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles             and JavaScript.<br \/>Advertisement<br \/>                              <a data-test=\"journal-link\" href=\"\/srep\"><i data-test=\"journal-title\">Scientific Reports<\/i><\/a>                          <b data-test=\"journal-volume\"><span class=\"u-visually-hidden\">volume<\/span>\u00a012<\/b>, Article\u00a0number:\u00a0<span data-test=\"article-number\">13635<\/span> (<span data-test=\"article-publication-year\">2022<\/span>)             <a href=\"#citeas\" class=\"c-article-info-details__cite-as u-hide-print\" data-track=\"click\" data-track-action=\"cite this article\" data-track-label=\"link\">Cite this article<\/a>                     <br \/><a href=\"\/articles\/s41598-022-17818-4\/metrics\" data-track=\"click\" data-track-action=\"view metrics\" data-track-label=\"link\" rel=\"nofollow\">Metrics <span class=\"u-visually-hidden\">details<\/span><\/a><br \/>Image segmentation by thresholding is an important and fundamental task in image processing and computer vision. In this paper, a new bi-level thresholding approach based on weighted Parzen-window and linear programming techniques is proposed to use in image thresholding segmentation. First, by proposing a weighted Parzen-window to describe the gray level distribution status, we obtain the boundaries for the foreground and background of the image. Then the image thresholding problem can be transformed into the problem of solving a linear programming problem for computing the coefficient values of weighted Parzen-window. The results of testing on synthetic, NDT and a set of benchmark images indicate that the proposed method can achieve a higher segmentation accuracy and robustness in comparison to some classical thresholding methods, such as inter class variance method (OTSU), Kapur\u2019s entropy-based method (KSW), and some state-of-art methods that consider spatial information, such as CHPSO, GLLV histogram method and GABOR histogram method.<br \/>Image segmentation technology has been widely applied in industry, agriculture, military fields, etc. Among all the image segmentation methods, thresholding is one of the most widely used, because it is simple and ease to implement. Image thresholding methods are divided into two categories, one is bi-level thresholding methods and the other is multi-level methods. The bi-level methods that involve the fundamental assumption that the foregrounds and backgrounds of the image have different gray level distributions, segment the image to foregrounds and backgrounds. The multi-level methods can be generalized by the bi-level methods that segment the image to multiple non-overlapping regions<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 1\" title=\"Lei, B. &amp; Fan, J. Image thresholding segmentation method based on minimum square rough entropy. Appl. Soft Comput. J. 84(11), 105687 (2019).\" href=\"\/articles\/s41598-022-17818-4#ref-CR1\" id=\"ref-link-section-d48721197e421\">1<\/a><\/sup>. So far, many successful thresholding methods have been developed and applied in many fields, such as infrared nondestructive testing, magnetic resonance imaging, etc.<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 2\" title=\"Sezgin, M. &amp; Sankur, B. Survey over image thresholding techniques and quantitative performance evaluation. J. Electron. Imaging 13(1), 146\u2013165 (2004).\" href=\"\/articles\/s41598-022-17818-4#ref-CR2\" id=\"ref-link-section-d48721197e425\">2<\/a><\/sup>. In this paper, we focus on bi-level thresholding methods.<br \/>In the process of bi-level thresholding, it is assumed that there exists an optimal threshold value separating the gray levels. The segmentation task can be implemented by classifying pixels whose gray level is less than the threshold as backgrounds and pixels whose gray level is greater than the threshold as foregrounds, or vice versa. For decades, some classical bi-level thresholding algorithms have been proposed, such as the inter-class variance method (OTSU)<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 3\" title=\"Otsu, N. A threshold selection method from gray-level histograms. IEEE Trans. Syst. Man Cybern. 9(1), 62\u201366 (1979).\" href=\"\/articles\/s41598-022-17818-4#ref-CR3\" id=\"ref-link-section-d48721197e432\">3<\/a><\/sup>, minimum error bi-level thresholding method (MET)<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 4\" title=\"Kittiler, J. &amp; Illingworth, J. Minimum error thresholding. Pattern Recogn. 19(1), 41\u201347 (1986).\" href=\"\/articles\/s41598-022-17818-4#ref-CR4\" id=\"ref-link-section-d48721197e436\">4<\/a><\/sup>, the entropic bi-level thresholding method based on one-dimensional histogram (1D KSW)<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 5\" title=\"Kapur, J. N., Sahoo, P. K. &amp; Wong, A. A new method for gray-level picture thresholding using the entropy of the histogram. Comput. Vis. Graph. Image Process. 29(3), 273\u2013285 (1985).\" href=\"\/articles\/s41598-022-17818-4#ref-CR5\" id=\"ref-link-section-d48721197e440\">5<\/a><\/sup>, Renyi\u2019s entropic bi-level thresholding method<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 6\" title=\"Sahoo, P. K., Wilkins, C. &amp; Yeager, J. Threshold selection using Renyi\u2019s entropy. Pattern Recogn. 30(1), 71\u201384 (1997).\" href=\"\/articles\/s41598-022-17818-4#ref-CR6\" id=\"ref-link-section-d48721197e444\">6<\/a><\/sup> and Tsallis\u2019s entropic bi-level thresholding method<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 7\" title=\"Portes, M., Esquef, I. A. &amp; Gesualdi, A. R. Image thresholding using Tsallis entropy. Pattern Recogn. Lett. 25(9), 1059\u20131065 (2004).\" href=\"\/articles\/s41598-022-17818-4#ref-CR7\" id=\"ref-link-section-d48721197e448\">7<\/a><\/sup> etc. In addition, these classical methods have been modified or combined with other techniques to develop numerous successful bi-level or multi-level thresholding methods. As typical examples, relative entropy theory and 3D histogram were combined with MET for an optimal threshold discriminant<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 8\" title=\"Liu, J., Zheng, J., Tang, Q. &amp; Jin, W. Minimum error thresholding segmentation algorithm based on 3D grayscale histogram. Math. Probl. Eng. 2014(pt.1), 1\u201313 (2014).\" href=\"\/articles\/s41598-022-17818-4#ref-CR8\" id=\"ref-link-section-d48721197e453\">8<\/a><\/sup>. The ant colony optimization approach was combined with the inter-class variance method for fast find out multiple thresholds of the images<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 9\" title=\"Qin, J., Shen, X., Mei, F. &amp; Fang, Z. An Otsu multi-thresholds segmentation algorithm based on improved ACO. J. Supercomput. 75, 955\u2013967 (2019).\" href=\"\/articles\/s41598-022-17818-4#ref-CR9\" id=\"ref-link-section-d48721197e457\">9<\/a><\/sup>. The hybrid whale optimization approach was combined with the 1D KSW method for multi-level thresholding segmentation<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 10\" title=\"Lang, C. &amp; Jia, H. Kapur\u2019s entropy for color image segmentation based on a hybrid Whale optimization algorithm. Entropy 21(3), 318 (2019).\" href=\"\/articles\/s41598-022-17818-4#ref-CR10\" id=\"ref-link-section-d48721197e461\">10<\/a>,<a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 11\" title=\"Zheping, Y., Jinzhong, Z., Zewen, Y. &amp; Jialing, T. Kapur\u2019s entropy for underwater multilevel thresholding image segmentation based on whale optimization algorithm. IEEE Access 9, 41429\u201341319 (2021).\" href=\"\/articles\/s41598-022-17818-4#ref-CR11\" id=\"ref-link-section-d48721197e464\">11<\/a><\/sup>. The meta-heuristics approach was combined with Renyi\u2019s entropy-based method for multi-level thresholding segmentation<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 12\" title=\"Liu, W. et al. Renyi\u2019s entropy based multilevel thresholding using a novel meta-heuristics algorithm. Appl. Sci. 10(9), 3225 (2020).\" href=\"\/articles\/s41598-022-17818-4#ref-CR12\" id=\"ref-link-section-d48721197e468\">12<\/a><\/sup>. The particle swarm optimization approach was combined with the Tsallis entropy-based method for multi-level thresholding segmentation<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 13\" title=\"Borjigin, S. &amp; Sahoo, P. K. Color image segmentation based on multi-level Tsallis\u2013Havrda\u2013Charv\u00e1t entropy and 2D histogram using PSO algorithms. Pattern Recogn. 92, 107\u2013118 (2019).\" href=\"\/articles\/s41598-022-17818-4#ref-CR13\" id=\"ref-link-section-d48721197e472\">13<\/a><\/sup>. The convergence heterogeneous particle swarm optimization algorithm, was utilized to find the optimal bi-level and multi-level thresholds<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 14\" title=\"Mozaffari, M. H. &amp; Lee, W. S. Convergent heterogeneous particle swarm optimisation algorithm for multilevel image thresholding segmentation. IET Image Proc. 11(8), 605\u2013619 (2017).\" href=\"\/articles\/s41598-022-17818-4#ref-CR14\" id=\"ref-link-section-d48721197e476\">14<\/a><\/sup>. The coyote optimization algorithm, which takes Ostu and fuzzy entropy as objective functions, was used to multi-level thresholds selection<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 15\" title=\"LinGuo, L., Lijuan, S., Yu, X., Shujing, L. &amp; Romany, F. M. Fuzzy multilevel image thresholding based on improved coyote optimization algorithm. IEEE Access 9, 33595\u201333607 (2021).\" href=\"\/articles\/s41598-022-17818-4#ref-CR15\" id=\"ref-link-section-d48721197e481\">15<\/a><\/sup>. Although many thresholding methods have been developed, the entropy-based methods remain the most popular. Many extensions of the entropy-based method, which are based on 1D histogram, have been proposed in recent years. For examples, Cheng proposed a new bi-level thresholding method by implementing fuzzy segmentation based on two-dimensional (2D) histograms<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 16\" title=\"Cheng, H. &amp; Chen, Y. Fuzzy partition of two-dimensional histogram and its application to thresholding. Pattern Recogn. 32, 825\u2013843 (1999).\" href=\"\/articles\/s41598-022-17818-4#ref-CR16\" id=\"ref-link-section-d48721197e485\">16<\/a><\/sup>. Xiao proposed two new entropic bi-level thresholding methods. The first method employs gray level spatial correlation (GLSC) histogram<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 17\" title=\"Xiao, Y., Cao, Z. &amp; Zhong, S. New entropic thresholding approach using gray-level spatial correlation histogram. Opt. Eng. 49, 1127\u20131134 (2010).\" href=\"\/articles\/s41598-022-17818-4#ref-CR17\" id=\"ref-link-section-d48721197e489\">17<\/a><\/sup>. In contrast to the 2D histogram, the GLSC histogram is obtained using the gray level of the pixels and their neighbors with similar gray level. The second method employs gray level and gradient magnitude (GLGM) histogram<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 18\" title=\"Xiao, Y., Cao, Z. &amp; Zhong, S. Entropic image thresholding based on GLGM histogram. Pattern Recogn. Lett. 40, 47\u201355 (2014).\" href=\"\/articles\/s41598-022-17818-4#ref-CR18\" id=\"ref-link-section-d48721197e493\">18<\/a><\/sup>. The GLGM histogram clearly captures the occurrence probability and spatial distribution features of gray level at the same time, and considers spatial information. Utilizing the orientation histogram of a gradient image to calculate the local edge property, a new bi-level thresholding method employing 2D-D histogram was proposed by Yimit<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 19\" title=\"Yimit, A., Hagihara, Y., Miyoshi, T. &amp; Hagihara, Y. 2-D direction histogram based entropic thresholding. Neuro Comput. 120, 287\u2013297 (2013).\" href=\"\/articles\/s41598-022-17818-4#ref-CR19\" id=\"ref-link-section-d48721197e497\">19<\/a><\/sup>. A new thresholding method based on a GLLV histogram was proposed by Zheng<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 20\" title=\"Zheng, X. L., Ye, H. &amp; Tang, Y. G. Image bi-level thresholding based on gray level-local variance histogram. Entropy 19, 191 (2017).\" href=\"\/articles\/s41598-022-17818-4#ref-CR20\" id=\"ref-link-section-d48721197e501\">20<\/a><\/sup> using the gray level information of pixels and its local variance in a neighborhood. A new thresholding method based on a GABOR histogram was proposed by Yi<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 21\" title=\"Yi, S., Zhang, G., He, J. &amp; Tong, L. Entropic image thresholding segmentation based on Gabor histogram. KSII Trans. Internet Inf. Syst. 13(4), 2113\u20132128 (2019).\" href=\"\/articles\/s41598-022-17818-4#ref-CR21\" id=\"ref-link-section-d48721197e506\">21<\/a><\/sup>. Recently, Xiong et al. proposed a new image thresholding method combining Kapur\u2019s entropy with Parzen-window estimation<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 22\" title=\"Fusong, X., Jian, Z., Yun, L. &amp; Zhiqiang, Z. (2021). A novel image thresholding method combining entropy with Parzen window estimation. Comput. J., bxab182, 2021, November.\" href=\"\/articles\/s41598-022-17818-4#ref-CR22\" id=\"ref-link-section-d48721197e510\">22<\/a><\/sup>.<br \/>In general, the improved 2D histogram methods outperform 1D histogram methods. However, the 2D entropic thresholding methods still have some limitations, such as, not a generic method for image thresholding, and lack of robustness or stability etc.<br \/>In this paper, we try to propose a new bi-level thresholding method, which is based on the boundaries for the foreground and background by using a weighted Parzen-window to describe the gray level distribution status rather than gray level probability density distributions (1D or 2D histogram) for the foreground and background in an image. Subsequently, image thresholding\u00a0was successfully transformed into a linear programming problem.\u00a0We used\u00a0the\u00a0simplex method to solve the linear programming problem. In the experimental section, the proposed method is compared with the classic and state-of-art methods to demonstrate its accuracy and robustness. The novel contribution of this study is the construction of a new data distribution description method based on the weighted Parzen-window technique, which can be regarded as a linear programming problem. This process is illustrated in Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig1\">1<\/a>.<br \/>Process of the proposed thresholding method.<br \/>The rest parts of this paper is organized as follows. In section\u00a0&#8220;<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"section anchor\" href=\"\/articles\/s41598-022-17818-4#Sec2\">The proposed method<\/a>&#8220;, we briefly introduce the Parzen-window technology, and provide a new bi-level thresholding method based on the weighted Parzen-window and linear programming. In section\u00a0&#8220;<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"section anchor\" href=\"\/articles\/s41598-022-17818-4#Sec5\">Experimental results<\/a>&#8220;, the results of the experiments and a discussion are presented. Finally, section\u00a0&#8220;<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"section anchor\" href=\"\/articles\/s41598-022-17818-4#Sec12\">Conclusions<\/a>&#8221; gives the conclusion.<br \/>For a gray image <span class=\"mathjax-tex\">(F=left{fleft(x,yright)|xin left{mathrm{1,2},3,dots ,mright}, yin left{mathrm{1,2},3,dots ,nright}right})<\/span> of size <span class=\"mathjax-tex\">(mtimes n)<\/span> with <span class=\"mathjax-tex\">(L)<\/span> gray levels, the gray level set <span class=\"mathjax-tex\">(G=left{mathrm{0,1},2,dots ,L-1right})<\/span><i>.<\/i> <span class=\"mathjax-tex\">(fleft(x,yright)in G)<\/span> is the gray value of the pixel located at location <span class=\"mathjax-tex\">(left(mathrm{x},mathrm{y}right))<\/span>. Let <span class=\"mathjax-tex\">({omega }_{l}=left{left(x,yright) |fleft(x,yright)=l, xin left{mathrm{1,2},3,dots ,mright},yin left{mathrm{1,2},3,dots ,nright}, lin Gright})<\/span>,<span class=\"mathjax-tex\">({C}_{l} left(lin Gright))<\/span> represents the number of pixels in <span class=\"mathjax-tex\">({omega }_{l})<\/span> , then <span class=\"mathjax-tex\">(omega =left{ {omega }_{l}, lin Gright})<\/span> and <span class=\"mathjax-tex\">(N=sum_{l=0}^{L-1}{C}_{l})<\/span> . Obviously, <span class=\"mathjax-tex\">({omega }_{l})<\/span> is defined in 2D space. Suppose that <span class=\"mathjax-tex\">(t)<\/span> is the threshold value, the result of bi-level thresholding by <span class=\"mathjax-tex\">(mathrm{t})<\/span> is a binary function <span class=\"mathjax-tex\">({f}_{t}left(x,yright))<\/span>:<br \/>The result of thresholding is a clustering problem that separates all pixels into two classes <span class=\"mathjax-tex\">(O)<\/span> and <span class=\"mathjax-tex\">(B)<\/span>. Where <span class=\"mathjax-tex\">(O)<\/span> represents foregrounds and <span class=\"mathjax-tex\">(B)<\/span> represents backgrounds, or vice versa.<br \/>Traditional thresholding methods are first used to compute the probability of each gray level distribution. Then, the optimal threshold value was computed by optimizing an appropriate objective function, which was designed using the gray level distribution or other properties.<br \/>As we known well, the Parzen-window estimation is an effective non-parametric estimation with solid theoretical foundation, which can better describe the distributions of data<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" title=\"Bian, Z. &amp; Zhang, X. Pattern Recognition 2nd edn. (Tsinghua University Press, 2000).\" href=\"#ref-CR23\" id=\"ref-link-section-d48721197e1398\">23<\/a>,<a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" title=\"Duda, R. O., Hart, P. E. &amp; Stork, D. G. (2003). Pattern Classification (H.D. Li, T.X, Yao, Trans.), Machinery Industry Press.\" href=\"#ref-CR24\" id=\"ref-link-section-d48721197e1398_1\">24<\/a>,<a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 25\" title=\"Torkkola, K. Feature extraction by non-parametric mutual information maximization. J. Mach. Learn. Res. 3, 1415\u20131438 (2003).\" href=\"\/articles\/s41598-022-17818-4#ref-CR25\" id=\"ref-link-section-d48721197e1401\">25<\/a><\/sup>. The basic idea is to estimate the <span class=\"mathjax-tex\">(pdf)<\/span> using the mean value of the densities of each point within a certain range. If we want to estimate the <span class=\"mathjax-tex\">(pdf)<\/span> at point <span class=\"mathjax-tex\">(X)<\/span>, we can place a window of size <span class=\"mathjax-tex\">(h)<\/span> at <span class=\"mathjax-tex\">(X)<\/span> and see how many observations of <span class=\"mathjax-tex\">({X}_{i})<\/span> fall into this window. The value of <span class=\"mathjax-tex\">(pdf)<\/span> is the average of the observations falling into this window. The Parzen-window estimate <span class=\"mathjax-tex\">({P}_{n}left(Xright))<\/span> can be expressed as:<br \/>where <span class=\"mathjax-tex\">({V}_{n})<\/span> is the volume of the d-D hypercube with edge length <span class=\"mathjax-tex\">({h}_{n})<\/span>, <span class=\"mathjax-tex\">({V}_{n}={h}_{n}^{d})<\/span>, <span class=\"mathjax-tex\">({h}_{n}=frac{c}{sqrt{n}})<\/span> is called window width. <span class=\"mathjax-tex\">(mathrm{c})<\/span> is a constant parameter that always takes the value of 1. <span class=\"mathjax-tex\">(Kleft(cdot right))<\/span> is the d-D kernel function (window function), and:<br \/>The most commonly used kernel function is the Gaussian kernel function (normal distribution), defined as:<br \/>Following the Parzen-window estimation, for the 2-D image <span class=\"mathjax-tex\">(F)<\/span>, the sample <span class=\"mathjax-tex\">(left(x,yright))<\/span> in the two-dimensional point space <span class=\"mathjax-tex\">({omega }_{l})<\/span>, its <span class=\"mathjax-tex\">(pdf)<\/span> <span class=\"mathjax-tex\">(pleft(x,{y,omega }_{l}right))<\/span> can be estimated by Eqs.\u00a0(<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"equation anchor\" href=\"\/articles\/s41598-022-17818-4#Equ5\">5<\/a>) and (<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"equation anchor\" href=\"\/articles\/s41598-022-17818-4#Equ6\">6<\/a>).<br \/>where <span class=\"mathjax-tex\">({C}_{l})<\/span> is the number of pixels in <span class=\"mathjax-tex\">({omega }_{l})<\/span>, <span class=\"mathjax-tex\">(pleft({omega }_{l}right))<\/span> can be approximated by a histogram, given by:<br \/>where <span class=\"mathjax-tex\">(N=sum_{l=0}^{L-1}{C}_{l})<\/span>. Then, the <span class=\"mathjax-tex\">(pleft(x,y,{omega }_{l}right))<\/span> is obtained by:<br \/>where <span class=\"mathjax-tex\">(left({x}_{j},{y}_{j}right))<\/span> denotes the coordinates of <span class=\"mathjax-tex\">(jth)<\/span> sample (pixel) in <span class=\"mathjax-tex\">({omega }_{l})<\/span>, <span class=\"mathjax-tex\">({V}_{{C}_{l}})<\/span> represents the volume of the cube whose edge length is <span class=\"mathjax-tex\">({sigma }_{l})<\/span>, <span class=\"mathjax-tex\">({sigma }_{l})<\/span> is also called the window width, that is, for a two-dimensional image, <span class=\"mathjax-tex\">({V}_{{C}_{l}}={{sigma }_{l}}^{2})<\/span>. <span class=\"mathjax-tex\">(varphi left(cdot right))<\/span> is a window function (also called kernel function). We chose Gaussian kernel function, which is taken as:<br \/>Thus, <span class=\"mathjax-tex\">(pleft(x,yright))<\/span> can be estimated using:<br \/>However, probability density estimation itself is an ill-posed problem. Moreover, the estimation of the probability density function involves a large amount of calculation, and it is easily affected by noise and the number of samples. To avoid these negative effects, we give up the estimation of probability density. We attempt to obtain the boundaries for the foregrounds and backgrounds in an image by using a weighted Parzen-window, to obtain a good description of the gray level distribution status, the thresholding problem can be converted to the problem of solving a linear programming problem for determining the coefficient values of the weighted Parzen-window.<br \/>Here, we propose the weighted Parzen-window method, which is an improvement of the Parzen-window method. By combining the proposed weighted Parzen-window method and a linear programming technique, we provide a new image thresholding method.<br \/>As is well know, thresholding segmentation assumes that the pixels are divided into two classes <span class=\"mathjax-tex\">({varvec{O}})<\/span> and <span class=\"mathjax-tex\">({varvec{B}})<\/span>. If we can choose an suitable <span class=\"mathjax-tex\">(rho)<\/span> to divide <span class=\"mathjax-tex\">(left{{omega }_{l}, lin G, G=left{mathrm{1,2},dots ,.L-1right}right})<\/span> into two classes, such as <span class=\"mathjax-tex\">(left{{omega }_{O}, Oin Gright})<\/span>, <span class=\"mathjax-tex\">(left{{omega }_{B},Bin Gright})<\/span>, <span class=\"mathjax-tex\">({omega }_{O}bigcap {omega }_{B}=phi , {omega }_{O}bigcup {omega }_{B}=G)<\/span>, and satisfying:<br \/>Then, Eq.\u00a0(<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"equation anchor\" href=\"\/articles\/s41598-022-17818-4#Equ13\">13<\/a>) can be seen as the boundary of foregrounds and backgrounds:<br \/>According to the boundary, it is easy to divide the gray level into two classes. However, this approach is not always effective, because the Parzen-window technique does not provide a method for choosing an appropriate <span class=\"mathjax-tex\">(uprho)<\/span>. Thus, the Parzen-window technique must be modified so that it can better describe the boundary of the data distribution and obtain the appropriate <span class=\"mathjax-tex\">(uprho)<\/span>. We now provide a solution strategy.<br \/>Suppose that a d-D pattern space with <span class=\"mathjax-tex\">(N)<\/span> samples is as follows:<br \/>where <span class=\"mathjax-tex\">(I)<\/span> denotes the coordinate set. Now, let\u2019s consider the following linear programming (LP).<br \/>where <span class=\"mathjax-tex\">(varphi left(cdot right))<\/span> denotes the kernel function. Because the kernel function <span class=\"mathjax-tex\">(varphi left(cdot right))<\/span> and coefficient <span class=\"mathjax-tex\">({a}_{i})<\/span> are non-negative, the implicit constraint is <span class=\"mathjax-tex\">(rho ge 0)<\/span>. The solution of the above LP and the kernel function together constitute a new description of the data distribution:<br \/>where <span class=\"mathjax-tex\">({I}^{^{prime}}=left{i|iin I and {a}_{i}&gt;0right})<\/span>. The solution of Eq.\u00a0(<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"equation anchor\" href=\"\/articles\/s41598-022-17818-4#Equ15\">15<\/a>) is guaranteed by the following theorem.<br \/>The solution of Eq.\u00a0(<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"equation anchor\" href=\"\/articles\/s41598-022-17818-4#Equ15\">15<\/a>) is absolute existence.<br \/>According to the constraints in Eq.\u00a0(<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"equation anchor\" href=\"\/articles\/s41598-022-17818-4#Equ15\">15<\/a>), we have:<br \/>It is easy to know that Eq.\u00a0(<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"equation anchor\" href=\"\/articles\/s41598-022-17818-4#Equ15\">15<\/a>) is a feasible solution.<br \/>Thus, Eq.\u00a0(<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"equation anchor\" href=\"\/articles\/s41598-022-17818-4#Equ15\">15<\/a>) exist in the solution domain. It can be concluded from the optimization theory that the solution of Eq.\u00a0(<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"equation anchor\" href=\"\/articles\/s41598-022-17818-4#Equ15\">15<\/a>) is an absolute existence. Proof end.<br \/>If <span class=\"mathjax-tex\">(varphi left({X}_{j},{X}_{i}right))<\/span> is regarded as a measure of the similarity between samples <span class=\"mathjax-tex\">(j)<\/span> and <span class=\"mathjax-tex\">(i)<\/span>, Eq.\u00a0(<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"equation anchor\" href=\"\/articles\/s41598-022-17818-4#Equ15\">15<\/a>) provides a strategy for selecting <span class=\"mathjax-tex\">(uprho)<\/span>. The constraints of Eq.\u00a0(<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"equation anchor\" href=\"\/articles\/s41598-022-17818-4#Equ15\">15<\/a>) make the inter-class similarity as large as possible. Therefore, the boundaries of the data distribution can be better delineated. Simultaneously, <span class=\"mathjax-tex\">(widehat{p}left(Xright))<\/span> is not the probability density estimation, instead, focus on describing the boundaries of data distribution, and:<br \/>The simplex method<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 26\" title=\"Dantzig, G. Linear Programming and Extensions (Princeton Univ. Press, 2016).\" href=\"\/articles\/s41598-022-17818-4#ref-CR26\" id=\"ref-link-section-d48721197e5311\">26<\/a><\/sup> is the most commonly used method to solve the LP problem, thus, we chose it for this study.<br \/>A gray image is regarded as a two-dimensional sample space. This can be easily mapped to linear programming. For example, a 2-dimensional space <span class=\"mathjax-tex\">(X)<\/span> is replaced by <span class=\"mathjax-tex\">(left{fleft(x,yright)|xin left{mathrm{1,2},3,dots ,mright},yin left{mathrm{1,2},3,dots ,nright}right})<\/span> , the index coordinate <span class=\"mathjax-tex\">(I)<\/span> is replaced by the pixel coordinate set <span class=\"mathjax-tex\">(omega)<\/span>, kernel function <span class=\"mathjax-tex\">(varphi left(cdot right))<\/span> is the same as in Eq.\u00a0(<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"equation anchor\" href=\"\/articles\/s41598-022-17818-4#Equ9\">9<\/a>). Thus, we can classify all gray levels into two classes using the proposed weighted Parzen-window and linear programming based image thresholding (WPWLPT) method.<br \/>In this section, we present the experimental results, obtained by using some classic methods (such as OTSU<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 3\" title=\"Otsu, N. A threshold selection method from gray-level histograms. IEEE Trans. Syst. Man Cybern. 9(1), 62\u201366 (1979).\" href=\"\/articles\/s41598-022-17818-4#ref-CR3\" id=\"ref-link-section-d48721197e5511\">3<\/a><\/sup> and KSW<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 5\" title=\"Kapur, J. N., Sahoo, P. K. &amp; Wong, A. A new method for gray-level picture thresholding using the entropy of the histogram. Comput. Vis. Graph. Image Process. 29(3), 273\u2013285 (1985).\" href=\"\/articles\/s41598-022-17818-4#ref-CR5\" id=\"ref-link-section-d48721197e5515\">5<\/a><\/sup>), some state-of-art methods (such as CHPSO<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 14\" title=\"Mozaffari, M. H. &amp; Lee, W. S. Convergent heterogeneous particle swarm optimisation algorithm for multilevel image thresholding segmentation. IET Image Proc. 11(8), 605\u2013619 (2017).\" href=\"\/articles\/s41598-022-17818-4#ref-CR14\" id=\"ref-link-section-d48721197e5519\">14<\/a><\/sup>, GLLV<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 19\" title=\"Yimit, A., Hagihara, Y., Miyoshi, T. &amp; Hagihara, Y. 2-D direction histogram based entropic thresholding. Neuro Comput. 120, 287\u2013297 (2013).\" href=\"\/articles\/s41598-022-17818-4#ref-CR19\" id=\"ref-link-section-d48721197e5523\">19<\/a><\/sup> and GABOR<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 20\" title=\"Zheng, X. L., Ye, H. &amp; Tang, Y. G. Image bi-level thresholding based on gray level-local variance histogram. Entropy 19, 191 (2017).\" href=\"\/articles\/s41598-022-17818-4#ref-CR20\" id=\"ref-link-section-d48721197e5527\">20<\/a><\/sup>, and all the parameters being set to the default values during the experiments) and our proposed method (we call it as WPWLPT from now). Li et al.<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 14\" title=\"Mozaffari, M. H. &amp; Lee, W. S. Convergent heterogeneous particle swarm optimisation algorithm for multilevel image thresholding segmentation. IET Image Proc. 11(8), 605\u2013619 (2017).\" href=\"\/articles\/s41598-022-17818-4#ref-CR14\" id=\"ref-link-section-d48721197e5532\">14<\/a><\/sup> proposed the CHPSO method, which can be used for both bi-level and multi-level thresholding (they provide equations for both the bi-level and multi-level cases). It uses OTSU and KAPUR as objective functions, which we denote CHPSO_otsu and CHPSO_ksw, respectively.<br \/>In order to assess the effectiveness of the proposed method, we qualitatively and quantitatively assessed on lots of images. For brevity, we only reported 22 representative thresholding results, which included two synthetic, eight nondestructive testing (NDT) and a set of benchmark images. These images had different sizes and histogram types. We designed two synthetic images. The NDT images were obtained from<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 2\" title=\"Sezgin, M. &amp; Sankur, B. Survey over image thresholding techniques and quantitative performance evaluation. J. Electron. Imaging 13(1), 146\u2013165 (2004).\" href=\"\/articles\/s41598-022-17818-4#ref-CR2\" id=\"ref-link-section-d48721197e5539\">2<\/a><\/sup>. The benchmark images belong to the Image Processing Standard Database (<a href=\"http:\/\/www.imageprocessingplace.com\/root_files_V3\/image_databases.htm\">http:\/\/www.imageprocessingplace.com\/root_files_V3\/image_databases.htm<\/a>) and the USC-SIPI Image Database (<a href=\"http:\/\/sipi.usc.edu\/database\/\">http:\/\/sipi.usc.edu\/database\/<\/a>), which are well-known and widely used in the image thresholding literatures.<br \/>Currently, there are several measurements<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 2\" title=\"Sezgin, M. &amp; Sankur, B. Survey over image thresholding techniques and quantitative performance evaluation. J. Electron. Imaging 13(1), 146\u2013165 (2004).\" href=\"\/articles\/s41598-022-17818-4#ref-CR2\" id=\"ref-link-section-d48721197e5560\">2<\/a>,<a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 21\" title=\"Yi, S., Zhang, G., He, J. &amp; Tong, L. Entropic image thresholding segmentation based on Gabor histogram. KSII Trans. Internet Inf. Syst. 13(4), 2113\u20132128 (2019).\" href=\"\/articles\/s41598-022-17818-4#ref-CR21\" id=\"ref-link-section-d48721197e5563\">21<\/a>,<a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" title=\"Zhang, Y. J. A survey on evaluation methods for image segmentation. Pattern Recogn. 29(8), 1335\u20131346 (1996).\" href=\"#ref-CR27\" id=\"ref-link-section-d48721197e5566\">27<\/a>,<a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" title=\"Roldan, R. R. et al. A measure of quality for evaluating methods of segmentation and edge detection. Pattern Recogn. 34(5), 969\u2013980 (2001).\" href=\"#ref-CR28\" id=\"ref-link-section-d48721197e5566_1\">28<\/a>,<a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 29\" title=\"Zhang, L., Mou, X. &amp; Zhang, D. FSIM: A feature similarity index for image quality assessment. IEEE Trans. Image Process. 20(8), 2378\u20132386 (2011).\" href=\"\/articles\/s41598-022-17818-4#ref-CR29\" id=\"ref-link-section-d48721197e5569\">29<\/a>,<a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 32\" title=\"Bowen, C., Ishan, M., Alexander G. S., Alexander, K. &amp; Rohit, G. (2022). Masked-attention mask transformer for universal image segmentation. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition. Last revised 15 Jun 2022.\" href=\"\/articles\/s41598-022-17818-4#ref-CR32\" id=\"ref-link-section-d48721197e5572\">32<\/a><\/sup> to quantitatively evaluate the quality of the image thresholding method. We used the misclassification error (<span class=\"mathjax-tex\">(ME)<\/span>)<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 2\" title=\"Sezgin, M. &amp; Sankur, B. Survey over image thresholding techniques and quantitative performance evaluation. J. Electron. Imaging 13(1), 146\u2013165 (2004).\" href=\"\/articles\/s41598-022-17818-4#ref-CR2\" id=\"ref-link-section-d48721197e5597\">2<\/a><\/sup>, region nonuniformity (<span class=\"mathjax-tex\">(NU)<\/span>)<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 2\" title=\"Sezgin, M. &amp; Sankur, B. Survey over image thresholding techniques and quantitative performance evaluation. J. Electron. Imaging 13(1), 146\u2013165 (2004).\" href=\"\/articles\/s41598-022-17818-4#ref-CR2\" id=\"ref-link-section-d48721197e5622\">2<\/a><\/sup>, feature similarity (<span class=\"mathjax-tex\">(FSIM)<\/span>)<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 29\" title=\"Zhang, L., Mou, X. &amp; Zhang, D. FSIM: A feature similarity index for image quality assessment. IEEE Trans. Image Process. 20(8), 2378\u20132386 (2011).\" href=\"\/articles\/s41598-022-17818-4#ref-CR29\" id=\"ref-link-section-d48721197e5648\">29<\/a><\/sup> and mean intersection over union (<span class=\"mathjax-tex\">(mIoU)<\/span>)<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 32\" title=\"Bowen, C., Ishan, M., Alexander G. S., Alexander, K. &amp; Rohit, G. (2022). Masked-attention mask transformer for universal image segmentation. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition. Last revised 15 Jun 2022.\" href=\"\/articles\/s41598-022-17818-4#ref-CR32\" id=\"ref-link-section-d48721197e5673\">32<\/a><\/sup> to quantitatively assess the different thresholding methods.<br \/><span class=\"mathjax-tex\">(ME)<\/span> measurement reflects the incorrect classification of foregrounds pixels to the backgrounds or vice versa<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 2\" title=\"Sezgin, M. &amp; Sankur, B. Survey over image thresholding techniques and quantitative performance evaluation. J. Electron. Imaging 13(1), 146\u2013165 (2004).\" href=\"\/articles\/s41598-022-17818-4#ref-CR2\" id=\"ref-link-section-d48721197e5700\">2<\/a><\/sup>. For the bi-level image thresholding problem, <span class=\"mathjax-tex\">(ME)<\/span> can be taken as:<br \/>where <span class=\"mathjax-tex\">({B}_{mathrm{o}})<\/span> and <span class=\"mathjax-tex\">({F}_{mathrm{o}})<\/span> denote the backgrounds and foregrounds of the optimal thresholded image, <span class=\"mathjax-tex\">({B}_{T})<\/span> and <span class=\"mathjax-tex\">({F}_{T})<\/span> denote the backgrounds and foregrounds region pixels of the original image, and <span class=\"mathjax-tex\">(left|*right|)<\/span> denotes the cardinality of the set <span class=\"mathjax-tex\">(*)<\/span>. Obviously, <span class=\"mathjax-tex\">(ME)<\/span> equals to 1 for the worst case and 0 for the best case. <span class=\"mathjax-tex\">(ME)<\/span> is the easiest and most effective method for the discrepancy measure<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 30\" title=\"Yasnoff, W. A., Mui, J. K. &amp; Bacus, J. W. Error measures for scence segmentation. Pattern Recogn. 9(4), 217\u2013231 (1997).\" href=\"\/articles\/s41598-022-17818-4#ref-CR30\" id=\"ref-link-section-d48721197e6005\">30<\/a><\/sup>.<br \/><span class=\"mathjax-tex\">(NU)<\/span> measures the intrinsic quality of the segmented regions, is defined as:<br \/>where <span class=\"mathjax-tex\">({sigma }^{2})<\/span> denotes the variance of the image, and <span class=\"mathjax-tex\">({{sigma }_{f}}^{2})<\/span> denotes the variance of the foregrounds. <span class=\"mathjax-tex\">({B}_{T})<\/span> and <span class=\"mathjax-tex\">({F}_{T})<\/span> denote the backgrounds and foregrounds region pixels of the original image. Obviously, <span class=\"mathjax-tex\">(NU)<\/span> closes to 0 for a well-segmented image and equals to 1 for the worst-segmented image.<br \/><span class=\"mathjax-tex\">(FSIM)<\/span> calculates the similarity of two images, is defined as:<br \/>where<br \/>where <span class=\"mathjax-tex\">({T}_{1})<\/span> and <span class=\"mathjax-tex\">({T}_{2})<\/span> denote constants. Here, <span class=\"mathjax-tex\">({T}_{1}=0.85, {T}_{2}=160)<\/span>. <span class=\"mathjax-tex\">(Omega)<\/span> is the whole space of image. <span class=\"mathjax-tex\">(G)<\/span> is the gradient of image, defined as:<br \/><span class=\"mathjax-tex\">(PC)<\/span> represents the phase consistency, defined as:<br \/>where <span class=\"mathjax-tex\">({A}_{n}left(xright))<\/span> denotes <span class=\"mathjax-tex\">(n)<\/span> order amplitude, <span class=\"mathjax-tex\">(Eleft(Xright))<\/span> represents <span class=\"mathjax-tex\">(n)<\/span> order response vector level at position <span class=\"mathjax-tex\">(X)<\/span>. <span class=\"mathjax-tex\">(varepsilon)<\/span> represents a small positive constant. Obviously, <span class=\"mathjax-tex\">(FSIM)<\/span> closes to 1 for a well-segmented result and equals to 0 for the worst-segmented result.<br \/>Synthetic images are perfect for testing the image thresholding algorithm because their optimal threshold values can be obtained manually<sup><a data-track=\"click\" data-track-action=\"reference anchor\" data-track-label=\"link\" data-test=\"citation-ref\" aria-label=\"Reference 31\" title=\"Bazi, Y., Bruzzone, L. &amp; Melgani, F. Image thresholding based on the EM algorithm and generalized Gaussian distribution. Pattern Recogn. 40, 619\u2013634 (2007).\" href=\"\/articles\/s41598-022-17818-4#ref-CR31\" id=\"ref-link-section-d48721197e7131\">31<\/a><\/sup>. Figure\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a> shows two original synthetic images with <span class=\"mathjax-tex\">(256times 256)<\/span> pixels, which named as \u201cCircles\u201d and \u201cSquares\u201d [Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>a,e], respectively. In Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>a, we place some circles (their gray level is 150) on a darker background (gray level is 50). Figure\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>b shows a noisy image of Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>a with Gaussian noise, and Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>c shows the histogram of Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>b. Figure\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>d shows the ground-truth image of Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>b. In Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>e, we place some squares (their gray level is 225) on a darker background (gray level is 75). Figure\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>f shows a noisy image of Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>e with Gaussian noise, and Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>g shows the histogram of Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>f. Figure\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>h shows the ground-truth image of Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig2\">2<\/a>f.<br \/>Two examples of synthetic images. (<b>a<\/b>) \u201cCircles\u201d image. (<b>b<\/b>) Noised \u201cCircles\u201d image. (<b>c<\/b>) Histogram of the noised \u201cCircles\u201d image. (<b>d<\/b>) Ground-truth image of noised \u201cCircles\u201d. (<b>e<\/b>) \u201cSquares\u201d image. (<b>f<\/b>) Noised \u201cSquares\u201d image. (<b>g<\/b>) Histogram of the noised \u201cSquares\u201d image. (<b>h<\/b>) Ground-truth image of noised \u201cSquares\u201d.<br \/>For the synthetic \u201cCircles\u201d image, the optimal threshold value, which was calculated manually based on the ground-truth image, is 108. The threshold values, <span class=\"mathjax-tex\">(MEs)<\/span>, <span class=\"mathjax-tex\">(NUs)<\/span> and <span class=\"mathjax-tex\">(FSIMs)<\/span> obatined using the seven thresholding methods are listed in Table <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab1\">1<\/a>. The best values in terms of <span class=\"mathjax-tex\">(MEs)<\/span>, <span class=\"mathjax-tex\">(NUs)<\/span> and <span class=\"mathjax-tex\">(FSIMs)<\/span> are highlighted in bold. Among the seven thresholding methods, the threshold value obtained by WPWLPT is the closest to the optimal threshold value. It equals 110, and that the <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.0049, 0.0837 and 0.8103, respectively. The threshold value obtained using the OTSU method is 102. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.0120, 0.1025 and 0.7998, respectively. The threshold value obtained using the KSW method is 82. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.2650, 0.2208 and 0.6425, respectively. The threshold value obtained using the CHPSO_otsu method is 102. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.0120, 0.1025 and 0.7998, respectively. The threshold value obtained using the CHPSO_ksw method is 84. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.2595, 0.2172 and 0.6652, respectively. The threshold value obtained using the GLLV method is 101. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.0156, 0.1368 and 0.7921, respectively. And the threshold value obtained using the GABOR method is 105. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.0098, 0.1027 and 0.8024, respectively.<br \/>For the synthetic \u201cSquares\u201d image, the optimal threshold value, which was calculated manually based on the ground-truth image, is 153. The threshold values, <span class=\"mathjax-tex\">(MEs)<\/span>, <span class=\"mathjax-tex\">(NUs)<\/span> and <span class=\"mathjax-tex\">(FSIMs)<\/span> obatined using the seven thresholding methods are listed in Table <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab2\">2<\/a>. The best values in terms of <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are highlighted in bold. Among the seven thresholding methods, the threshold value obtained by WPWLPT was also the closest to the optimal threshold value. It equals 152, and that the <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.0018, 0.0386 and 0.8198, respectively. The threshold value obtained using the OTSU method is 147. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.0020, 0.0405 and 0.8002, respectively. The threshold value obtained using the KSW method is 108. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.0606, 0.1982 and 0.6512, respectively. The threshold value obtained using the CHPSO_otsu method is 148. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.0019, 0.0401 and 0.8076, respectively. The threshold value obtained using the CHPSO_ksw method is 110. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.0618, 0.1884 and 0.6725, respectively. The threshold value obtained using the GLLV method is 150. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.0018, 0.0398 and 0.8178, respectively. And the threshold value obtained using the GABOR method is 146. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> are equal to 0.0021, 0.0403 and 0.8104, respectively.<br \/>As can be seen from the results of the threshold values, <span class=\"mathjax-tex\">(MEs)<\/span>, <span class=\"mathjax-tex\">(NUs)<\/span> and <span class=\"mathjax-tex\">(FSIMs)<\/span>, it is clear that:<br \/>For the synthetic \u201cCircles\u201d image, the threshold values of KSW and CHPSO_ksw are 82 and 84, respectively. They are almost worthless threshold values, because of them far from the optimal threshold (108). In contrast, the threshold values of OTSU, CHPSO_otsu, GLLV, GABOR and WPWLPT are 102, 102, 101, 105, 110, respectively, which are reasonable threshold values due to their near to the optimal value. Especially, the threshold value of our WPWLPT is only 2 larger than the optimal threshold. The results in terms of <span class=\"mathjax-tex\">(MEs)<\/span>, <span class=\"mathjax-tex\">(NUs)<\/span> and <span class=\"mathjax-tex\">(FSIMs)<\/span> also reveal that our WPWLPT yields the best results. The <span class=\"mathjax-tex\">(MEs)<\/span> and <span class=\"mathjax-tex\">(NUs)<\/span> provided by the KSW and CHPSO_ksw methods were so higher, and the <span class=\"mathjax-tex\">(FSIMs)<\/span> were so lower than other methods. While OTSU, CHPSO_otsu, GLLV, GABOR and WPWLPT can obtain reasonable results, especially our WPWLPT method which obtains the minimum <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and the maximum <span class=\"mathjax-tex\">(FSIM)<\/span> values.<br \/>For the synthetic \u201cSquares\u201d image, the threshold values of KSW and CHPSO_ksw are 108 and 110, respectively. These are almost worthless threshold values too, because they are from the optimal threshold (153). In contrast, the threshold values of OTSU, CHPSO_otsu, GLLV, GABOR and WPWLPT are 147, 148, 150, 146, 152, respectively, which are reasonable threshold values because they are close to the optimal value. Especially, the threshold value of our WPWLPT is only 1 less than the optimal threshold. The results in terms of <span class=\"mathjax-tex\">(MEs)<\/span>, <span class=\"mathjax-tex\">(NUs)<\/span> and <span class=\"mathjax-tex\">(FSIMs)<\/span> also reveal that our WPWLPT yields the minimum <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and the maximum <span class=\"mathjax-tex\">(FSIM)<\/span> values, which were best results among all the seven thresholding methods.<br \/>Figure\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig3\">3<\/a> provides a visual comparison between the thresholding results obtained by the OTSU, KSW, CHPSO_otsu, CHPSO_ksw, GLLV, GABOR and the proposed WPWLPT methods. As can be seen from Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig3\">3<\/a>, the KSW and CHPSO_ksw methods obtained almost unvalued results because their segmented images had obvious noise (see Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig3\">3<\/a>, the second and fourth images of each row). In contrast, all the OTSU, CHPSO_otsu, GLLV, GABOR and WPWLPT methods segmented a cleaner image because the threshold values they obtained were close to the optimal threshold value. Furthermore, by zooming in Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig3\">3<\/a>, we can easily observe that the WPWLPT method gives the clearest segmentation results compared with the OTSU, CHPSO_otsu, GLLV and GABOR methods, because it has the least residual noise.<br \/>Thresholding segmentation results of the noised synthetic image by using different methods. For each experimental synthetic image, the results of OTSU, KSW, CHPSO_otsu, CHPSO_ksw, GLLV, GABOR and WPWLPT methods are displayed side by side from left to right.<br \/>The NDT images are also ideal for testing the image thresholding algorithm because their ground-truth images can be obtained directly. In this part, eight NDT images were used to assess the performance of the WPWLPT. They are \u201cPCB\u201d, \u201cdefective tile\u201d, \u201cmaterial structure\u201d, \u201cfuselage material\u201d, \u201ceddy current\u201d, \u201cultrasonic\u201d, \u201cGFRP\u201dand \u201cbonemarr\u201d. All the above eight images, their histograms and ground-truth images are shown in Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig4\">4<\/a>. The thresholding segmentation results obtained using the reference thresholding methods and WPWLPT are shown in Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig5\">5<\/a>.<br \/>The experimental NDT images, their histograms and ground-truth images.<br \/>Thresholding segmentation results by using different methods. For each experimental image, the results of OTSU, KSW, CHPSO_otsu, CHPSO_ksw, GLLV, GABOR and WPWLPT methods are displayed side by side from left to right.<br \/>Tables <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab3\">3<\/a>, <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab4\">4<\/a> and <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab5\">5<\/a> show the <span class=\"mathjax-tex\">(MEs)<\/span>, <span class=\"mathjax-tex\">(NUs)<\/span> and <span class=\"mathjax-tex\">(FSIMs)<\/span> of the different thresholding methods, respectively. <span class=\"mathjax-tex\">({varpi }_{ME})<\/span>,<span class=\"mathjax-tex\">({varpi }_{NU})<\/span> and <span class=\"mathjax-tex\">({varpi }_{FSIM})<\/span> represent the average of <span class=\"mathjax-tex\">(MEs)<\/span>, <span class=\"mathjax-tex\">(NUs)<\/span> and <span class=\"mathjax-tex\">(FSIMs)<\/span> respectively. The best results are highlighted in bold.<br \/>Because of length limits, we only analyzed four NDT images. They are the \u201cPCB\u201d, \u201cmaterial structure\u201d, \u201ceddy current\u201d and \u201cGFRP\u201d.<br \/>For \u201cPCB\u201d image, the <span class=\"mathjax-tex\">(ME)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> values obtained by WPWLPT method are optimal, while the <span class=\"mathjax-tex\">(NU)<\/span> value is inferior to GLLV method only. This is reasonable because they focused on different aspects of measurement. <span class=\"mathjax-tex\">(ME)<\/span> represents the percentage of background pixels incorrectly classified to the foreground, or vice versa, <span class=\"mathjax-tex\">(FSIM)<\/span> focuses on the texture, shape and other features, while <span class=\"mathjax-tex\">(NU)<\/span> judges the intrinsic quality of the segmented areas. The worst results are obtained using the KSW method. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> values are equal to 0.2170, 0.6725 and 0.5864, respectively. Obviously, compared with other methods, its <span class=\"mathjax-tex\">(ME)<\/span> and <span class=\"mathjax-tex\">(NU)<\/span> values are too high and <span class=\"mathjax-tex\">(FSIM)<\/span> value is too low, making the results worthless. A visual comparison, as shown in Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig5\">5<\/a>, shows that the OTSU, CHPSO_otsu, GLLV and WPWLPT methods can segment better segmentation image. By comparison, the KSW, CHPSO_ksw and GABOR methods segmented valueless results because they misclassify lots of foregrounds as backgrounds (see Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig5\">5<\/a>, first row, second, fourth and sixth images, they can\u2019t distinguish the backgrounds and printed circuit board, especially the second image).<br \/>For \u201cmaterial structure\u201d image, GLLV, GABOR and OTSU yield the best <span class=\"mathjax-tex\">(ME)<\/span> <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> values, respectively. However, WPWLPT yields the closest values of <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> values to the best. The worst results are obtained using the KSW method. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> equal to 0.6176, 0.7036 and 0.5028, respectively. Obviously, its <span class=\"mathjax-tex\">(ME)<\/span> and <span class=\"mathjax-tex\">(NU)<\/span> values are too high and <span class=\"mathjax-tex\">(FSIM)<\/span> value is too low, so the results obtained are worthless. The visual comparison, as can be seen from Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig5\">5<\/a>, we can also discover that the OTSU, CHPSO_otsu, GLLV and WPWLPT methods can segment better segmentation image. By comparison, the KSW and CHPSO_ksw methods segment almost an unvalued segmentation image because they misclassify lots of foregrounds as backgrounds (see Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig5\">5<\/a>, third row, second and fourth image, the whole image looks black).<br \/>For \u201ceddy current\u201d image, the <span class=\"mathjax-tex\">(ME)<\/span> and <span class=\"mathjax-tex\">(NU)<\/span> values obtained by WPWLPT method are optimal, while the <span class=\"mathjax-tex\">(FISM)<\/span> value is inferior to GABOR method only. The worst results are obtained using the KSW method. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> values are equal to 0.0407, 0.1239 and 0.7002, respectively. Similar to the results of the KSW method, the CHPSO_ksw method also yielded poor results for <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> values. The visual comparison, as can be seen from Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig5\">5<\/a>, we can also discover that the GLLV, GABOR and WPWLPT methods can segment better segmentation images. By comparison, the KSW and CHPSO_ksw methods segment a low-value segmentation image because they misclassify some backgrounds as foregrounds (see Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig5\">5<\/a>, fifth row, second and fourth images, some black shadows appeared in the segmentation image).<br \/>For \u201cGFRP\u201d image, all the <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FISM)<\/span> values obtained by WPWLPT method are optimal. The worst results are also obtained by the KSW method. Its <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FSIM)<\/span> values are equal to 0.4898, 0.7879 and 0.6036, respectively. Obviously, the <span class=\"mathjax-tex\">(ME)<\/span> (0.7879!) is close to 1 which corresponding to the worst case. The results obtained by CHPSO_ksw method are also valueless due to their higher <span class=\"mathjax-tex\">(ME)<\/span>, <span class=\"mathjax-tex\">(NU)<\/span> values and lower <span class=\"mathjax-tex\">(FSIM)<\/span> values. A visual comparison, as can be seen from Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig5\">5<\/a>, shows that the OTSU, CHPAO_otsu, GLLV, GABOR and WPWLPT methods can segment better segmentation images. By comparison, the KSW and CHPSO_ksw methods segmented valueless results because they misclassify lots of backgrounds as foregrounds (see Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig5\">5<\/a>, seventh row, second and fourth images, it&#8217;s impossible to distinguish between foregrounds and backgrounds).<br \/>A set of benchmark images belonging to the Image Processing Standard Database and USC-SIPI Image Database, which contain 12\u00a0Gy images. For brevity, we give 12 images here, \u201ccameraman\u201d, \u201chouse\u201d, \u201cjetplane\u201d, \u201clake\u201d, \u201cmilkdrop\u201d, \u201clivingroom\u201d, \u201cmandril\u201d, \u201cpeppers\u201d, \u201cpirate\u201d, \u201cwalkbridge\u201d, \u201ctank\u201d, and \u201cboat\u201d, all in uncompressed tif or tiff format and of the same <span class=\"mathjax-tex\">(512times 512)<\/span> size. The thresholding segmentation results of the corresponding twelve images obtained by the reference thresholding methods and WPWLPT are shown row by row from top to bottom in Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig6\">6<\/a>. Tables <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab6\">6<\/a> and <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab7\">7<\/a> show the <span class=\"mathjax-tex\">(NUs)<\/span>, and <span class=\"mathjax-tex\">(FSIMs)<\/span> of different thresholding methods, respectively. <span class=\"mathjax-tex\">({varpi }_{NU})<\/span>, and <span class=\"mathjax-tex\">({varpi }_{FSIM})<\/span> represent the average of <span class=\"mathjax-tex\">(MEs)<\/span>, and <span class=\"mathjax-tex\">(FSIMs)<\/span> respectively. The best results are highlighted in bold. It should be noted that, we did not employ <span class=\"mathjax-tex\">(ME)<\/span> to measure the quality of thresholding methods experiment on these images, because the ideal thresholded or ground-truth images cannot be acquired.<br \/>Thresholding segmentation results by using different methods. For each experimental image, the results of OTSU, KSW, CHPSO_otsu, CHPSO_ksw, GLLV, GABOR and WPWLPT methods are displayed side by side from left to right.<br \/>As shown in Fig.\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig6\">6<\/a>, Tables <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab6\">6<\/a> and <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab7\">7<\/a>, it can be observed that:<br \/>For most of the tested images, the values of <span class=\"mathjax-tex\">(NU)<\/span> and <span class=\"mathjax-tex\">(FISM)<\/span> obtained by WPWLPT are the lowest. Specifically, for the \u201ccameraman\u201d, \u201chouse\u201d, \u201cmilkdrop\u201d, \u201cpeppers\u201d, \u201cwalkbridge\u201d, \u201ctank\u201d and \u201cboat\u201d images, the proposed WPWLPT method can obtain the lowest <span class=\"mathjax-tex\">(NU)<\/span> values. For the \u201cpirate\u201d image, OTSU obtains the lowest <span class=\"mathjax-tex\">(NU)<\/span> values. For the \u201clake\u201d and \u201cpirate\u201d images, CHPSO_otsu obtains the lowest <span class=\"mathjax-tex\">(NU)<\/span> values. For the \u201clivingroom\u201d and \u201cmandril\u201d images, GLLV obtains the lowest <span class=\"mathjax-tex\">(NU)<\/span> values. For the \u201cjetplane\u201d image, GABOR obtains the lowest <span class=\"mathjax-tex\">(NU)<\/span> values. In terms of <span class=\"mathjax-tex\">(FISM)<\/span>, the proposed method are similar. It obtains the highest <span class=\"mathjax-tex\">(FISM)<\/span> values in the \u201ccameraman\u201d, \u201cmilkdrop\u201d, \u201cpeppers\u201d, \u201cpirate\u201d, \u201cwalkbridge\u201d, \u201ctank\u201d and \u201cboat\u201d images. Although the WPWLPT method did not obtain the lowest <span class=\"mathjax-tex\">(NU)<\/span> values and highest <span class=\"mathjax-tex\">(FISM)<\/span> values for all 12 test images, the average value of <span class=\"mathjax-tex\">(NUs)<\/span> and <span class=\"mathjax-tex\">(FISMs)<\/span> were the best among all seven thresholding methods. Specifically, the <span class=\"mathjax-tex\">({varpi }_{NU})<\/span> values of OTSU, KSW, CHPSO_otsu, CHPSO_ksw, GLLV, GABOR and WPWLPT are equal to 0.1119, 0.3772, 0.1127, 0.3686, 0.2037, 0.1278 and 0.0992, respectively. The <span class=\"mathjax-tex\">({varpi }_{FISM})<\/span> values of OTSU, KSW, CHPSO_otsu, CHPSO_ksw, GLLV, GABOR and WPWLPT equal to 0.7601, 0.5896, 0.7597, 0.6058, 0.73014, 0.7646 and 0.7867, respectively. It means that the <span class=\"mathjax-tex\">({varpi }_{NU})<\/span> value obtain by our method outperforms the OTSU, KSW, CHPSO_otsu, CHPSO_ksw, GLLV and GABOR methods by 1.27%, 27.80%, 1.35%, 26.94%, 10.45% and 2.87% respectively, while the <span class=\"mathjax-tex\">({varpi }_{FISM})<\/span> value outperforms the OTSU, KSW, CHPSO_otsu, CHPSO_ksw, GLLV and GABOR methods by 2.66%, 19.71%, 2.69%, 18.09%, 5.66% and 2.20% respectively.<br \/>From the above analysis, we can assert that the WPWLPT method can calculate better <span class=\"mathjax-tex\">({varpi }_{NU})<\/span> and <span class=\"mathjax-tex\">({varpi }_{FISM})<\/span> values in comparision with other reference methods. This result demonstrates the stability and accuracy of the proposed method.<br \/><span class=\"mathjax-tex\">(mIoU)<\/span> is the more widely used objective metric for the task of image segmentation. It is defined as:<br \/>where <span class=\"mathjax-tex\">(k)<\/span> is the number of classes, <span class=\"mathjax-tex\">(TP, FN and FP)<\/span> denote true positives, false positives and false positives, respectively. We employ <span class=\"mathjax-tex\">(mIoU)<\/span> to objectively evaluate synthetic and NDT images because they have ground-truth images. The <span class=\"mathjax-tex\">(mIoUs)<\/span> obtained by different thresholding methods are listed in Table <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab8\">8<\/a>. <span class=\"mathjax-tex\">({varpi }_{mIoU})<\/span> represents the average of <span class=\"mathjax-tex\">(mIoUs)<\/span>. The best results are highlighted in bold.<br \/>As shown in Table <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab8\">8<\/a>, in most test images, the values of <span class=\"mathjax-tex\">(mIoU)<\/span> obtained by WPWLPT are the highest. Specifically, for the \u201cCircles\u201d, \u201cPCB\u201d, \u201cdefective tile \u201c, \u201cmaterial structure\u201d, \u201cfuselage material \u201c, \u201ceddy current\u201d, \u201cultrasonic\u201d and \u201cGFRP\u201d images, the proposed WPWLPT method can obtain the highest <span class=\"mathjax-tex\">(mIoU)<\/span> values. For the \u201c, \u201cSquares\u201d image, GLLV, GABOR and WPWLPT both obtained the highest <span class=\"mathjax-tex\">(mIoU)<\/span> value. For the \u201cbonemarr\u201d image, GLLV and WPWLPT both got the highest <span class=\"mathjax-tex\">(mIoU)<\/span> values.<br \/>Figure\u00a0<a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig7\">7<\/a> depicts the average values of <span class=\"mathjax-tex\">(mIoU)<\/span> for the synthetic and NDT images. As shown in Fig. <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig7\">7<\/a>, our method has been improved to varying degrees on average <span class=\"mathjax-tex\">(mIoU)<\/span> compared with other methods. Specifically, the <span class=\"mathjax-tex\">({varpi }_{mIoU})<\/span> values of OTSU, KSW, CHPSO_otsu, CHPSO_ksw, GLLV, GABOR and WPWLPT are equal to 78.7%, 65.8%, 79.5%, 68.2%, 83.3%, 83.7% and 85.3%, respectively. It means that the <span class=\"mathjax-tex\">({varpi }_{mIoU})<\/span> value obtain by our method outperforms the OTSU, KSW, CHPSO_otsu, CHPSO_ksw, GLLV and GABOR methods by 6.0%, 19.5%, 5.8%, 17.1%, 2.0% and 1.6%, respectively.<br \/>The average <span class=\"mathjax-tex\">(mIoUs)<\/span> (%) of different methods.<br \/>Based on the above analysis of the experimental results of synthetic, NDT and a benchmark images, we find that:<br \/>The proposed WPWLPT method obtained the best segmentation performance for most images. In addition, our method can yields the lowest <span class=\"mathjax-tex\">({varpi }_{ME})<\/span> and <span class=\"mathjax-tex\">({varpi }_{NU})<\/span>, the highest <span class=\"mathjax-tex\">({varpi }_{FISM})<\/span> and <span class=\"mathjax-tex\">({varpi }_{mIoU})<\/span> on all the synthetic, NDT and the benchmark of images. Specifically, for the two synthetic images, the <span class=\"mathjax-tex\">({varpi }_{ME})<\/span>, <span class=\"mathjax-tex\">({varpi }_{NU})<\/span>, and <span class=\"mathjax-tex\">({varpi }_{FISM})<\/span> of WPWLPT equal to 0.0034, 0.0612 and 0.8151, respectively. The <span class=\"mathjax-tex\">({varpi }_{ME})<\/span> of WPWLPT outperforms the competing methods by 0.0037 to 0.1595, the <span class=\"mathjax-tex\">({varpi }_{NU})<\/span> outperforms the competing methods by 0.0102 to 0.1484, and the <span class=\"mathjax-tex\">({varpi }_{FISM})<\/span> outperforms the competing methods by 0.0101 to 0.1636. For the eight NDT image, the <span class=\"mathjax-tex\">({varpi }_{ME})<\/span>, <span class=\"mathjax-tex\">({varpi }_{NU})<\/span>, and <span class=\"mathjax-tex\">({varpi }_{FISM})<\/span> of WPWLPT equal to 0.0386, 0.0886 and 0.7168, respectively. The <span class=\"mathjax-tex\">({varpi }_{ME})<\/span> of WPWLPT outperforms the competing methods by 0.0146 to 0.2152, the <span class=\"mathjax-tex\">({varpi }_{NU})<\/span> outperforms the competing methods by 0.0088 to 0.2999, and the <span class=\"mathjax-tex\">({varpi }_{FISM})<\/span> outperforms the competing methods by 0.0192 to 0.0950. For the benchmark of twelve images, the <span class=\"mathjax-tex\">({varpi }_{NU})<\/span>, and <span class=\"mathjax-tex\">({varpi }_{FISM})<\/span> of WPWLPT equal to 0.0992, and 0.7867, respectively. The <span class=\"mathjax-tex\">({varpi }_{NU})<\/span> outperforms the competing methods by 0.0127 to 0.2780, and the <span class=\"mathjax-tex\">({varpi }_{FISM})<\/span> outperforms the competing methods by 0.0220 to 0.1971. For all the ten synthetic and NDT images, The <span class=\"mathjax-tex\">({varpi }_{mIoU})<\/span> of WPWLPT equals to 85.3%, which outperforms the competing methods by 1.6% to 19.5%. These experimental results well demonstrate the effectiveness and robustness of the proposed WPWLPT method.<br \/>From a visual perspective (Figs. <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig3\">3<\/a>, <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig5\">5<\/a> and <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"figure anchor\" href=\"\/articles\/s41598-022-17818-4#Fig6\">6<\/a>), although for some images, our method does not achieve the best segmentation effectiveness, it can obtain acceptable or close to the best results, which also shows the stability of our method.<br \/>OTSU is a traditional method, that exhibits high stability and accuracy. It outperformed most of the other methods except for ours.<br \/>Although our method works well for most images, it doesn\u2019t yield best performance on \u201cmaterial structure\u201d, \u201clake\u201d, \u201cmilkdrop\u201d et al. (For these three images, all the <span class=\"mathjax-tex\">({varpi }_{ME})<\/span>, <span class=\"mathjax-tex\">({varpi }_{NU})<\/span>, <span class=\"mathjax-tex\">({varpi }_{FISM})<\/span> or <span class=\"mathjax-tex\">({varpi }_{mIoU})<\/span> values are not optimal). The possible reason is the boundary information, which plays a crucial role in our proposed method is not obvious.<br \/>The KSW and CHPSO_ksw methods are the two worst performing methods. The Kapur based method is 1D entropy without considering other information. Obviously, GLLV and GABOR methods are superior to Kapur based method because they introduce other information such as gradient magnitude, texture and contour, etc. This also gave us inspiration to introduce other information into our method, which is our next step. This also gives us a clue to introduce other information into our method, and contour information is a potential choice. Moreover, it can be seen as our future work.<br \/>In addition to the qualitative and quantitative assessment, the running time of the thresholding method is another important evaluation. Table <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab9\">9<\/a> reports the average running time obtained by different threshold methods for the above 22 test images. All the experiments are running on the DELL notebook with Intel(R) Core (TM) i5-4300U CPU @ 1.90\u00a0GHz 2.50GHZ, 16\u00a0GB memory. The running environment is Matlab (R2015b). As shows in Table <a data-track=\"click\" data-track-label=\"link\" data-track-action=\"table anchor\" href=\"\/articles\/s41598-022-17818-4#Tab9\">9<\/a>, the proposed WPWLPT method takes approximately the same amount of time as the GLLV method. It is superior to the GABOR method but inferior to 1D methods, such as OTSU, KSW, CHPSO_otsu and CHPSO_ksw. Although our method is slower than OTSU, KSW etc., the running time is completely acceptable in many applications. In addition, our WPWLPT method is still highly competitive because of its superior effectiveness and robustness.<br \/>In this study, a new image bi-level thresholding method is proposed. The method first obtains the boundaries for the foreground and background in the image using a weighted Parzen-window to describe the gray level distribution status. Secondly, the image thresholding problem can be transformed into the problem of solving a linear programming problem for computing the coefficient values of the weighted Parzen-window. By solving the problem of linear programming, we determine the threshold. In the experiment, we used two synthetic, eight NDT and a benchmark of twelve testing images, which have different histogram types, to evaluate the quality of the proposed image thresholding method. The measurement of visual and quantitative results demonstrates that our proposed method, compared with the OTSU, KSW, CHPSO_otsu, CHPSO_ksw, GLLV and GABOR methods, can achieve better effectiveness and robustness. In the future, as an extension of this work, we will embed other information, such as texture, contour etc. in WPWLPT to enhance its performance, and extend the method to the problem of multilevel thresholding.<br \/>The data underlying this article will be shared on reasonable request to the corresponding author.<br \/>Lei, B. &amp; Fan, J. Image thresholding segmentation method based on minimum square rough entropy. <i>Appl. Soft Comput. J.<\/i> <b>84<\/b>(11), 105687 (2019).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1016%2Fj.asoc.2019.105687\" aria-label=\"Article reference 1\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 1\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Image%20thresholding%20segmentation%20method%20based%20on%20minimum%20square%20rough%20entropy&amp;journal=Appl.%20Soft%20Comput.%20J.&amp;volume=84&amp;issue=11&amp;publication_year=2019&amp;author=Lei%2CB&amp;author=Fan%2CJ\">                     Google Scholar<\/a>\u00a0                 <br \/>Sezgin, M. &amp; Sankur, B. Survey over image thresholding techniques and quantitative performance evaluation. <i>J. Electron. Imaging<\/i> <b>13<\/b>(1), 146\u2013165 (2004).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=2004JEI....13..146S\" aria-label=\"ADS reference 2\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1117%2F1.1631315\" aria-label=\"Article reference 2\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 2\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Survey%20over%20image%20thresholding%20techniques%20and%20quantitative%20performance%20evaluation&amp;journal=J.%20Electron.%20Imaging&amp;volume=13&amp;issue=1&amp;pages=146-165&amp;publication_year=2004&amp;author=Sezgin%2CM&amp;author=Sankur%2CB\">                     Google Scholar<\/a>\u00a0                 <br \/>Otsu, N. A threshold selection method from gray-level histograms. <i>IEEE Trans. Syst. Man Cybern.<\/i> <b>9<\/b>(1), 62\u201366 (1979).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1109%2FTSMC.1979.4310076\" aria-label=\"Article reference 3\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 3\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=A%20threshold%20selection%20method%20from%20gray-level%20histograms&amp;journal=IEEE%20Trans.%20Syst.%20Man%20Cybern.&amp;volume=9&amp;issue=1&amp;pages=62-66&amp;publication_year=1979&amp;author=Otsu%2CN\">                     Google Scholar<\/a>\u00a0                 <br \/>Kittiler, J. &amp; Illingworth, J. Minimum error thresholding. <i>Pattern Recogn.<\/i> <b>19<\/b>(1), 41\u201347 (1986).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=1986PatRe..19...41K\" aria-label=\"ADS reference 4\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1016%2F0031-3203%2886%2990030-0\" aria-label=\"Article reference 4\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 4\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Minimum%20error%20thresholding&amp;journal=Pattern%20Recogn.&amp;volume=19&amp;issue=1&amp;pages=41-47&amp;publication_year=1986&amp;author=Kittiler%2CJ&amp;author=Illingworth%2CJ\">                     Google Scholar<\/a>\u00a0                 <br \/>Kapur, J. N., Sahoo, P. K. &amp; Wong, A. A new method for gray-level picture thresholding using the entropy of the histogram. <i>Comput. Vis. Graph. Image Process.<\/i> <b>29<\/b>(3), 273\u2013285 (1985).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1016%2F0734-189X%2885%2990125-2\" aria-label=\"Article reference 5\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 5\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=A%20new%20method%20for%20gray-level%20picture%20thresholding%20using%20the%20entropy%20of%20the%20histogram&amp;journal=Comput.%20Vis.%20Graph.%20Image%20Process.&amp;volume=29&amp;issue=3&amp;pages=273-285&amp;publication_year=1985&amp;author=Kapur%2CJN&amp;author=Sahoo%2CPK&amp;author=Wong%2CA\">                     Google Scholar<\/a>\u00a0                 <br \/>Sahoo, P. K., Wilkins, C. &amp; Yeager, J. Threshold selection using Renyi\u2019s entropy. <i>Pattern Recogn.<\/i> <b>30<\/b>(1), 71\u201384 (1997).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=1997PatRe..30...71S\" aria-label=\"ADS reference 6\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1016%2FS0031-3203%2896%2900065-9\" aria-label=\"Article reference 6\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 6\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Threshold%20selection%20using%20Renyi%E2%80%99s%20entropy&amp;journal=Pattern%20Recogn.&amp;volume=30&amp;issue=1&amp;pages=71-84&amp;publication_year=1997&amp;author=Sahoo%2CPK&amp;author=Wilkins%2CC&amp;author=Yeager%2CJ\">                     Google Scholar<\/a>\u00a0                 <br \/>Portes, M., Esquef, I. A. &amp; Gesualdi, A. R. Image thresholding using Tsallis entropy. <i>Pattern Recogn. Lett.<\/i> <b>25<\/b>(9), 1059\u20131065 (2004).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=2004PaReL..25.1059P\" aria-label=\"ADS reference 7\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1016%2Fj.patrec.2004.03.003\" aria-label=\"Article reference 7\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 7\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Image%20thresholding%20using%20Tsallis%20entropy&amp;journal=Pattern%20Recogn.%20Lett.&amp;volume=25&amp;issue=9&amp;pages=1059-1065&amp;publication_year=2004&amp;author=Portes%2CM&amp;author=Esquef%2CIA&amp;author=Gesualdi%2CAR\">                     Google Scholar<\/a>\u00a0                 <br \/>Liu, J., Zheng, J., Tang, Q. &amp; Jin, W. Minimum error thresholding segmentation algorithm based on 3D grayscale histogram. <i>Math. Probl. Eng.<\/i> <b>2014<\/b>(pt.1), 1\u201313 (2014).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/www.ams.org\/mathscinet-getitem?mr=3166942\" aria-label=\"MathSciNet reference 8\">MathSciNet<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"\/articles\/cas-redirect\/1:CAS:528:DC%2BC2cXhsFWltb3J\" aria-label=\"CAS reference 8\">CAS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/www.emis.de\/MATH-item?1407.94017\" aria-label=\"MATH reference 8\">MATH<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 8\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Minimum%20error%20thresholding%20segmentation%20algorithm%20based%20on%203D%20grayscale%20histogram&amp;journal=Math.%20Probl.%20Eng.&amp;volume=2014&amp;issue=pt.1&amp;pages=1-13&amp;publication_year=2014&amp;author=Liu%2CJ&amp;author=Zheng%2CJ&amp;author=Tang%2CQ&amp;author=Jin%2CW\">                     Google Scholar<\/a>\u00a0                 <br \/>Qin, J., Shen, X., Mei, F. &amp; Fang, Z. An Otsu multi-thresholds segmentation algorithm based on improved ACO. <i>J. Supercomput.<\/i> <b>75<\/b>, 955\u2013967 (2019).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1007%2Fs11227-018-2622-0\" aria-label=\"Article reference 9\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 9\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=An%20Otsu%20multi-thresholds%20segmentation%20algorithm%20based%20on%20improved%20ACO&amp;journal=J.%20Supercomput.&amp;volume=75&amp;pages=955-967&amp;publication_year=2019&amp;author=Qin%2CJ&amp;author=Shen%2CX&amp;author=Mei%2CF&amp;author=Fang%2CZ\">                     Google Scholar<\/a>\u00a0                 <br \/>Lang, C. &amp; Jia, H. Kapur\u2019s entropy for color image segmentation based on a hybrid Whale optimization algorithm. <i>Entropy<\/i> <b>21<\/b>(3), 318 (2019).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=2019Entrp..21..318L\" aria-label=\"ADS reference 10\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/www.ams.org\/mathscinet-getitem?mr=3943155\" aria-label=\"MathSciNet reference 10\">MathSciNet<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.3390%2Fe21030318\" aria-label=\"Article reference 10\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 10\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Kapur%E2%80%99s%20entropy%20for%20color%20image%20segmentation%20based%20on%20a%20hybrid%20Whale%20optimization%20algorithm&amp;journal=Entropy&amp;volume=21&amp;issue=3&amp;publication_year=2019&amp;author=Lang%2CC&amp;author=Jia%2CH\">                     Google Scholar<\/a>\u00a0                 <br \/>Zheping, Y., Jinzhong, Z., Zewen, Y. &amp; Jialing, T. Kapur\u2019s entropy for underwater multilevel thresholding image segmentation based on whale optimization algorithm. <i>IEEE Access<\/i> <b>9<\/b>, 41429\u201341319 (2021).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 11\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Kapur%E2%80%99s%20entropy%20for%20underwater%20multilevel%20thresholding%20image%20segmentation%20based%20on%20whale%20optimization%20algorithm&amp;journal=IEEE%20Access&amp;volume=9&amp;pages=41429-41319&amp;publication_year=2021&amp;author=Zheping%2CY&amp;author=Jinzhong%2CZ&amp;author=Zewen%2CY&amp;author=Jialing%2CT\">                     Google Scholar<\/a>\u00a0                 <br \/>Liu, W. <i>et al.<\/i> Renyi\u2019s entropy based multilevel thresholding using a novel meta-heuristics algorithm. <i>Appl. Sci.<\/i> <b>10<\/b>(9), 3225 (2020).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"\/articles\/cas-redirect\/1:CAS:528:DC%2BB3cXhtlGgtLbO\" aria-label=\"CAS reference 12\">CAS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.3390%2Fapp10093225\" aria-label=\"Article reference 12\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 12\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Renyi%27s%20entropy%20based%20multilevel%20thresholding%20using%20a%20novel%20meta-heuristics%20algorithm&amp;journal=Appl.%20Sci.&amp;volume=10&amp;issue=9&amp;publication_year=2020&amp;author=Liu%2CW&amp;author=Huang%2CY&amp;author=Ye%2CZ&amp;author=Cai%2CW&amp;author=Cheng%2CX&amp;author=Ibrahim%2CF\">                     Google Scholar<\/a>\u00a0                 <br \/>Borjigin, S. &amp; Sahoo, P. K. Color image segmentation based on multi-level Tsallis\u2013Havrda\u2013Charv\u00e1t entropy and 2D histogram using PSO algorithms. <i>Pattern Recogn.<\/i> <b>92<\/b>, 107\u2013118 (2019).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=2019PatRe..92..107B\" aria-label=\"ADS reference 13\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1016%2Fj.patcog.2019.03.011\" aria-label=\"Article reference 13\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 13\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Color%20image%20segmentation%20based%20on%20multi-level%20Tsallis%E2%80%93Havrda%E2%80%93Charv%C3%A1t%20entropy%20and%202D%20histogram%20using%20PSO%20algorithms&amp;journal=Pattern%20Recogn.&amp;volume=92&amp;pages=107-118&amp;publication_year=2019&amp;author=Borjigin%2CS&amp;author=Sahoo%2CPK\">                     Google Scholar<\/a>\u00a0                 <br \/>Mozaffari, M. H. &amp; Lee, W. S. Convergent heterogeneous particle swarm optimisation algorithm for multilevel image thresholding segmentation. <i>IET Image Proc.<\/i> <b>11<\/b>(8), 605\u2013619 (2017).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1049%2Fiet-ipr.2016.0489\" aria-label=\"Article reference 14\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 14\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Convergent%20heterogeneous%20particle%20swarm%20optimisation%20algorithm%20for%20multilevel%20image%20thresholding%20segmentation&amp;journal=IET%20Image%20Proc.&amp;volume=11&amp;issue=8&amp;pages=605-619&amp;publication_year=2017&amp;author=Mozaffari%2CMH&amp;author=Lee%2CWS\">                     Google Scholar<\/a>\u00a0                 <br \/>LinGuo, L., Lijuan, S., Yu, X., Shujing, L. &amp; Romany, F. M. Fuzzy multilevel image thresholding based on improved coyote optimization algorithm. <i>IEEE Access<\/i> <b>9<\/b>, 33595\u201333607 (2021).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1109%2FACCESS.2021.3060749\" aria-label=\"Article reference 15\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 15\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Fuzzy%20multilevel%20image%20thresholding%20based%20on%20improved%20coyote%20optimization%20algorithm&amp;journal=IEEE%20Access&amp;volume=9&amp;pages=33595-33607&amp;publication_year=2021&amp;author=LinGuo%2CL&amp;author=Lijuan%2CS&amp;author=Yu%2CX&amp;author=Shujing%2CL&amp;author=Romany%2CFM\">                     Google Scholar<\/a>\u00a0                 <br \/>Cheng, H. &amp; Chen, Y. Fuzzy partition of two-dimensional histogram and its application to thresholding. <i>Pattern Recogn.<\/i> <b>32<\/b>, 825\u2013843 (1999).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=1999PatRe..32..825C\" aria-label=\"ADS reference 16\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"\/articles\/cas-redirect\/1:CAS:528:DyaK1MXhsFSkt7g%3D\" aria-label=\"CAS reference 16\">CAS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1016%2FS0031-3203%2898%2900080-6\" aria-label=\"Article reference 16\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 16\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Fuzzy%20partition%20of%20two-dimensional%20histogram%20and%20its%20application%20to%20thresholding&amp;journal=Pattern%20Recogn.&amp;volume=32&amp;pages=825-843&amp;publication_year=1999&amp;author=Cheng%2CH&amp;author=Chen%2CY\">                     Google Scholar<\/a>\u00a0                 <br \/>Xiao, Y., Cao, Z. &amp; Zhong, S. New entropic thresholding approach using gray-level spatial correlation histogram. <i>Opt. Eng.<\/i> <b>49<\/b>, 1127\u20131134 (2010).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1117%2F1.3526333\" aria-label=\"Article reference 17\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 17\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=New%20entropic%20thresholding%20approach%20using%20gray-level%20spatial%20correlation%20histogram&amp;journal=Opt.%20Eng.&amp;volume=49&amp;pages=1127-1134&amp;publication_year=2010&amp;author=Xiao%2CY&amp;author=Cao%2CZ&amp;author=Zhong%2CS\">                     Google Scholar<\/a>\u00a0                 <br \/>Xiao, Y., Cao, Z. &amp; Zhong, S. Entropic image thresholding based on GLGM histogram. <i>Pattern Recogn. Lett.<\/i> <b>40<\/b>, 47\u201355 (2014).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=2014PaReL..40...47X\" aria-label=\"ADS reference 18\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1016%2Fj.patrec.2013.12.017\" aria-label=\"Article reference 18\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 18\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Entropic%20image%20thresholding%20based%20on%20GLGM%20histogram&amp;journal=Pattern%20Recogn.%20Lett.&amp;volume=40&amp;pages=47-55&amp;publication_year=2014&amp;author=Xiao%2CY&amp;author=Cao%2CZ&amp;author=Zhong%2CS\">                     Google Scholar<\/a>\u00a0                 <br \/>Yimit, A., Hagihara, Y., Miyoshi, T. &amp; Hagihara, Y. 2-D direction histogram based entropic thresholding. <i>Neuro Comput.<\/i> <b>120<\/b>, 287\u2013297 (2013).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 19\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=2-D%20direction%20histogram%20based%20entropic%20thresholding&amp;journal=Neuro%20Comput.&amp;volume=120&amp;pages=287-297&amp;publication_year=2013&amp;author=Yimit%2CA&amp;author=Hagihara%2CY&amp;author=Miyoshi%2CT&amp;author=Hagihara%2CY\">                     Google Scholar<\/a>\u00a0                 <br \/>Zheng, X. L., Ye, H. &amp; Tang, Y. G. Image bi-level thresholding based on gray level-local variance histogram. <i>Entropy<\/i> <b>19<\/b>, 191 (2017).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=2017Entrp..19..191Z\" aria-label=\"ADS reference 20\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.3390%2Fe19050191\" aria-label=\"Article reference 20\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 20\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Image%20bi-level%20thresholding%20based%20on%20gray%20level-local%20variance%20histogram&amp;journal=Entropy&amp;volume=19&amp;publication_year=2017&amp;author=Zheng%2CXL&amp;author=Ye%2CH&amp;author=Tang%2CYG\">                     Google Scholar<\/a>\u00a0                 <br \/>Yi, S., Zhang, G., He, J. &amp; Tong, L. Entropic image thresholding segmentation based on Gabor histogram. <i>KSII Trans. Internet Inf. Syst.<\/i> <b>13<\/b>(4), 2113\u20132128 (2019).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 21\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Entropic%20image%20thresholding%20segmentation%20based%20on%20Gabor%20histogram&amp;journal=KSII%20Trans.%20Internet%20Inf.%20Syst.&amp;volume=13&amp;issue=4&amp;pages=2113-2128&amp;publication_year=2019&amp;author=Yi%2CS&amp;author=Zhang%2CG&amp;author=He%2CJ&amp;author=Tong%2CL\">                     Google Scholar<\/a>\u00a0                 <br \/>Fusong, X., Jian, Z., Yun, L. &amp; Zhiqiang, Z. (2021). A novel image thresholding method combining entropy with Parzen window estimation. <i>Comput. J.<\/i>, bxab182, 2021, November.<br \/>Bian, Z. &amp; Zhang, X. <i>Pattern Recognition<\/i> 2nd edn. (Tsinghua University Press, 2000).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 23\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Pattern%20Recognition&amp;publication_year=2000&amp;author=Bian%2CZ&amp;author=Zhang%2CX\">                     Google Scholar<\/a>\u00a0                 <br \/>Duda, R. O., Hart, P. E. &amp; Stork, D. G. (2003). <i>Pattern Classification<\/i> (H.D. Li, T.X, Yao, Trans.), Machinery Industry Press.<br \/>Torkkola, K. Feature extraction by non-parametric mutual information maximization. <i>J. Mach. Learn. Res.<\/i> <b>3<\/b>, 1415\u20131438 (2003).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/www.ams.org\/mathscinet-getitem?mr=2020765\" aria-label=\"MathSciNet reference 25\">MathSciNet<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/www.emis.de\/MATH-item?1102.68638\" aria-label=\"MATH reference 25\">MATH<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 25\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Feature%20extraction%20by%20non-parametric%20mutual%20information%20maximization&amp;journal=J.%20Mach.%20Learn.%20Res.&amp;volume=3&amp;pages=1415-1438&amp;publication_year=2003&amp;author=Torkkola%2CK\">                     Google Scholar<\/a>\u00a0                 <br \/>Dantzig, G. <i>Linear Programming and Extensions<\/i> (Princeton Univ. Press, 2016).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 26\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Linear%20Programming%20and%20Extensions&amp;publication_year=2016&amp;author=Dantzig%2CG\">                     Google Scholar<\/a>\u00a0                 <br \/>Zhang, Y. J. A survey on evaluation methods for image segmentation. <i>Pattern Recogn.<\/i> <b>29<\/b>(8), 1335\u20131346 (1996).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=1996PatRe..29.1335Z\" aria-label=\"ADS reference 27\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1016%2F0031-3203%2895%2900169-7\" aria-label=\"Article reference 27\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 27\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=A%20survey%20on%20evaluation%20methods%20for%20image%20segmentation&amp;journal=Pattern%20Recogn.&amp;volume=29&amp;issue=8&amp;pages=1335-1346&amp;publication_year=1996&amp;author=Zhang%2CYJ\">                     Google Scholar<\/a>\u00a0                 <br \/>Roldan, R. R. <i>et al.<\/i> A measure of quality for evaluating methods of segmentation and edge detection. <i>Pattern Recogn.<\/i> <b>34<\/b>(5), 969\u2013980 (2001).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=2001PatRe..34..969R\" aria-label=\"ADS reference 28\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1016%2FS0031-3203%2800%2900052-2\" aria-label=\"Article reference 28\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 28\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=A%20measure%20of%20quality%20for%20evaluating%20methods%20of%20segmentation%20and%20edge%20detection&amp;journal=Pattern%20Recogn.&amp;volume=34&amp;issue=5&amp;pages=969-980&amp;publication_year=2001&amp;author=Roldan%2CRR&amp;author=Lopera%2CJFG&amp;author=Allah%2CCA\">                     Google Scholar<\/a>\u00a0                 <br \/>Zhang, L., Mou, X. &amp; Zhang, D. FSIM: A feature similarity index for image quality assessment. <i>IEEE Trans. Image Process.<\/i> <b>20<\/b>(8), 2378\u20132386 (2011).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=2011ITIP...20.2378L\" aria-label=\"ADS reference 29\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/www.ams.org\/mathscinet-getitem?mr=2866342\" aria-label=\"MathSciNet reference 29\">MathSciNet<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1109%2FTIP.2011.2109730\" aria-label=\"Article reference 29\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 29\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=FSIM%3A%20A%20feature%20similarity%20index%20for%20image%20quality%20assessment&amp;journal=IEEE%20Trans.%20Image%20Process.&amp;volume=20&amp;issue=8&amp;pages=2378-2386&amp;publication_year=2011&amp;author=Zhang%2CL&amp;author=Mou%2CX&amp;author=Zhang%2CD\">                     Google Scholar<\/a>\u00a0                 <br \/>Yasnoff, W. A., Mui, J. K. &amp; Bacus, J. W. Error measures for scence segmentation. <i>Pattern Recogn.<\/i> <b>9<\/b>(4), 217\u2013231 (1997).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=1977PatRe...9..217Y\" aria-label=\"ADS reference 30\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1016%2F0031-3203%2877%2990006-1\" aria-label=\"Article reference 30\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 30\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Error%20measures%20for%20scence%20segmentation&amp;journal=Pattern%20Recogn.&amp;volume=9&amp;issue=4&amp;pages=217-231&amp;publication_year=1997&amp;author=Yasnoff%2CWA&amp;author=Mui%2CJK&amp;author=Bacus%2CJW\">                     Google Scholar<\/a>\u00a0                 <br \/>Bazi, Y., Bruzzone, L. &amp; Melgani, F. Image thresholding based on the EM algorithm and generalized Gaussian distribution. <i>Pattern Recogn.<\/i> <b>40<\/b>, 619\u2013634 (2007).<br \/><a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"http:\/\/adsabs.harvard.edu\/cgi-bin\/nph-data_query?link_type=ABSTRACT&amp;bibcode=2007PatRe..40..619B\" aria-label=\"ADS reference 31\">ADS<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" href=\"https:\/\/doi.org\/10.1016%2Fj.patcog.2006.05.006\" aria-label=\"Article reference 31\">Article<\/a>\u00a0     <a data-track=\"click\" data-track-action=\"outbound reference\" data-track-label=\"link\" rel=\"nofollow noopener\" aria-label=\"Google Scholar reference 31\" href=\"http:\/\/scholar.google.com\/scholar_lookup?&amp;title=Image%20thresholding%20based%20on%20the%20EM%20algorithm%20and%20generalized%20Gaussian%20distribution&amp;journal=Pattern%20Recogn.&amp;volume=40&amp;pages=619-634&amp;publication_year=2007&amp;author=Bazi%2CY&amp;author=Bruzzone%2CL&amp;author=Melgani%2CF\">                     Google Scholar<\/a>\u00a0                 <br \/>Bowen, C., Ishan, M., Alexander G. S., Alexander, K. &amp; Rohit, G. (2022). Masked-attention mask transformer for universal image segmentation. <i>Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition<\/i>. Last revised 15 Jun 2022.<br \/><a data-track=\"click\" data-track-action=\"download citation references\" data-track-label=\"link\" href=\"https:\/\/citation-needed.springer.com\/v2\/references\/10.1038\/s41598-022-17818-4?format=refman&amp;flavour=references\">Download references<svg width=\"16\" height=\"16\" focusable=\"false\" role=\"img\" aria-hidden=\"true\" class=\"u-icon\"><use xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" xlink:href=\"#global-icon-download\"><\/use><\/svg><\/a><br \/>This work was partially supported by the National Natural Science Foundation of China (61672369, 6177255, 62072321 and 61972454), the Collaborative Innovation Center of Novel Software Technology and Industrialization, and the Priority Academic Program Development of Jiangsu Higher Education Institutions, China.<br \/>Soochow College, Soochow University, Suzhou, 215006, Jiangsu, China<br \/>Fusong Xiong,\u00a0Zhiqiang Zhang,\u00a0Yun Ling\u00a0&amp;\u00a0Jian Zhang<br \/>Collaborative Innovation Center of Novel Software Technology and Industrialization, Nanjing University, Nanjing, 210008, Jiangsu, China<br \/>Fusong Xiong,\u00a0Zhiqiang Zhang\u00a0&amp;\u00a0Yun Ling<br \/>Wenzheng College, Soochow University, Suzhou, 215104, Jiangsu, China<br \/>Jian Zhang<br \/>You can also search for this author in                         <span class=\"c-article-identifiers\"><a class=\"c-article-identifiers__item\" href=\"http:\/\/www.ncbi.nlm.nih.gov\/entrez\/query.fcgi?cmd=search&amp;term=Fusong%20Xiong\" data-track=\"click\" data-track-action=\"author link - pubmed\" data-track-label=\"link\" rel=\"nofollow\">PubMed<\/a><span class=\"u-hide\">\u00a0<\/span><a class=\"c-article-identifiers__item\" href=\"http:\/\/scholar.google.co.uk\/scholar?as_q=&amp;num=10&amp;btnG=Search+Scholar&amp;as_epq=&amp;as_oq=&amp;as_eq=&amp;as_occt=any&amp;as_sauthors=%22Fusong%20Xiong%22&amp;as_publication=&amp;as_ylo=&amp;as_yhi=&amp;as_allsubj=all&amp;hl=en\" data-track=\"click\" data-track-action=\"author link - scholar\" data-track-label=\"link\" rel=\"nofollow\">Google Scholar<\/a><\/span><br \/>You can also search for this author in                         <span class=\"c-article-identifiers\"><a class=\"c-article-identifiers__item\" href=\"http:\/\/www.ncbi.nlm.nih.gov\/entrez\/query.fcgi?cmd=search&amp;term=Zhiqiang%20Zhang\" data-track=\"click\" data-track-action=\"author link - pubmed\" data-track-label=\"link\" rel=\"nofollow\">PubMed<\/a><span class=\"u-hide\">\u00a0<\/span><a class=\"c-article-identifiers__item\" href=\"http:\/\/scholar.google.co.uk\/scholar?as_q=&amp;num=10&amp;btnG=Search+Scholar&amp;as_epq=&amp;as_oq=&amp;as_eq=&amp;as_occt=any&amp;as_sauthors=%22Zhiqiang%20Zhang%22&amp;as_publication=&amp;as_ylo=&amp;as_yhi=&amp;as_allsubj=all&amp;hl=en\" data-track=\"click\" data-track-action=\"author link - scholar\" data-track-label=\"link\" rel=\"nofollow\">Google Scholar<\/a><\/span><br \/>You can also search for this author in                         <span class=\"c-article-identifiers\"><a class=\"c-article-identifiers__item\" href=\"http:\/\/www.ncbi.nlm.nih.gov\/entrez\/query.fcgi?cmd=search&amp;term=Yun%20Ling\" data-track=\"click\" data-track-action=\"author link - pubmed\" data-track-label=\"link\" rel=\"nofollow\">PubMed<\/a><span class=\"u-hide\">\u00a0<\/span><a class=\"c-article-identifiers__item\" href=\"http:\/\/scholar.google.co.uk\/scholar?as_q=&amp;num=10&amp;btnG=Search+Scholar&amp;as_epq=&amp;as_oq=&amp;as_eq=&amp;as_occt=any&amp;as_sauthors=%22Yun%20Ling%22&amp;as_publication=&amp;as_ylo=&amp;as_yhi=&amp;as_allsubj=all&amp;hl=en\" data-track=\"click\" data-track-action=\"author link - scholar\" data-track-label=\"link\" rel=\"nofollow\">Google Scholar<\/a><\/span><br \/>You can also search for this author in                         <span class=\"c-article-identifiers\"><a class=\"c-article-identifiers__item\" href=\"http:\/\/www.ncbi.nlm.nih.gov\/entrez\/query.fcgi?cmd=search&amp;term=Jian%20Zhang\" data-track=\"click\" data-track-action=\"author link - pubmed\" data-track-label=\"link\" rel=\"nofollow\">PubMed<\/a><span class=\"u-hide\">\u00a0<\/span><a class=\"c-article-identifiers__item\" href=\"http:\/\/scholar.google.co.uk\/scholar?as_q=&amp;num=10&amp;btnG=Search+Scholar&amp;as_epq=&amp;as_oq=&amp;as_eq=&amp;as_occt=any&amp;as_sauthors=%22Jian%20Zhang%22&amp;as_publication=&amp;as_ylo=&amp;as_yhi=&amp;as_allsubj=all&amp;hl=en\" data-track=\"click\" data-track-action=\"author link - scholar\" data-track-label=\"link\" rel=\"nofollow\">Google Scholar<\/a><\/span><br \/>Conceptualization, F.X.; methodology, F.X. and J.Z.; software, F.X. and Z.Z.; validation, F.X. and J.Z.; formal analysis, Y.L.; writing\u2014review and editing, F.X. and J.Z.; All authors have read and agreed to the published version of the manuscript.<br \/>Correspondence to                 <a id=\"corresp-c1\" href=\"mailto:xiongfusong@suda.edu.cn\">Fusong Xiong<\/a> or <a id=\"corresp-c2\" href=\"mailto:zhangjian2012@suda.edu.cn\">Jian Zhang<\/a>.<br \/>The authors declare no competing interests.<br \/>Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.<br \/><b>Open Access<\/b>  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article&#8217;s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article&#8217;s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit <a href=\"http:\/\/creativecommons.org\/licenses\/by\/4.0\/\" rel=\"license\">http:\/\/creativecommons.org\/licenses\/by\/4.0\/<\/a>.<br \/><a data-track=\"click\" data-track-action=\"view rights and permissions\" data-track-label=\"link\" href=\"https:\/\/s100.copyright.com\/AppDispatchServlet?title=Image%20thresholding%20segmentation%20based%20on%20weighted%20Parzen-window%20and%20linear%20programming%20techniques&amp;author=Fusong%20Xiong%20et%20al&amp;contentID=10.1038%2Fs41598-022-17818-4&amp;copyright=The%20Author%28s%29&amp;publication=2045-2322&amp;publicationDate=2022-08-10&amp;publisherName=SpringerNature&amp;orderBeanReset=true&amp;oa=CC%20BY\">Reprints and Permissions<\/a><br \/>Xiong, F., Zhang, Z., Ling, Y. <i>et al.<\/i> Image thresholding segmentation based on weighted Parzen-window and linear programming techniques.                     <i>Sci Rep<\/i> <b>12, <\/b>13635 (2022). https:\/\/doi.org\/10.1038\/s41598-022-17818-4<br \/><a data-test=\"citation-link\" data-track=\"click\" data-track-action=\"download article citation\" data-track-label=\"link\" data-track-external=\"\" href=\"https:\/\/citation-needed.springer.com\/v2\/references\/10.1038\/s41598-022-17818-4?format=refman&amp;flavour=citation\">Download citation<svg width=\"16\" height=\"16\" focusable=\"false\" role=\"img\" aria-hidden=\"true\" class=\"u-icon\"><use xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" xlink:href=\"#global-icon-download\"><\/use><\/svg><\/a><br \/>Received<span class=\"u-hide\">: <\/span><span class=\"c-bibliographic-information__value\"><time datetime=\"2021-12-23\">23 December 2021<\/time><\/span><br \/>Accepted<span class=\"u-hide\">: <\/span><span class=\"c-bibliographic-information__value\"><time datetime=\"2022-08-01\">01 August 2022<\/time><\/span><br \/>Published<span class=\"u-hide\">: <\/span><span class=\"c-bibliographic-information__value\"><time datetime=\"2022-08-10\">10 August 2022<\/time><\/span><br \/><abbr title=\"Digital Object Identifier\">DOI<\/abbr><span class=\"u-hide\">: <\/span><span class=\"c-bibliographic-information__value\">https:\/\/doi.org\/10.1038\/s41598-022-17818-4<\/span><br \/>Anyone you share the following link with will be able to read this content:<br \/>Sorry, a shareable link is not currently available for this article.<\/p>\n<p>                             Provided by the Springer Nature SharedIt content-sharing initiative                         <br \/>By submitting a comment you agree to abide by our <a href=\"\/info\/tandc.html\">Terms<\/a> and <a href=\"\/info\/community-guidelines.html\">Community Guidelines<\/a>. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.<br \/>Advertisement<br \/>                 <a href=\"\/search\/advanced\"                    data-track=\"click\" data-track-action=\"advanced search\" data-track-label=\"link\">                     Advanced search                 <\/a>             <br \/>                 <span class=\"c-meta__item\">                     Scientific Reports (<i>Sci Rep<\/i>)                 <\/span>                           <span class=\"c-meta__item\">         <abbr title=\"International Standard Serial Number\">ISSN<\/abbr> <span itemprop=\"onlineIssn\">2045-2322<\/span> (online)     <\/span>                                           <br \/>&copy; 2022 Springer Nature Limited<br \/>Sign up for the <em>Nature Briefing<\/em> newsletter \u2014 what matters in science, free to your inbox daily.<\/p>\n<p><a href=\"https:\/\/www.nature.com\/articles\/s41598-022-17818-4\">source<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.Advertisement [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":62238,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-62237","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/loopbeatzz.info\/wp\/wp-json\/wp\/v2\/posts\/62237","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/loopbeatzz.info\/wp\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/loopbeatzz.info\/wp\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/loopbeatzz.info\/wp\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/loopbeatzz.info\/wp\/wp-json\/wp\/v2\/comments?post=62237"}],"version-history":[{"count":0,"href":"https:\/\/loopbeatzz.info\/wp\/wp-json\/wp\/v2\/posts\/62237\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/loopbeatzz.info\/wp\/wp-json\/wp\/v2\/media\/62238"}],"wp:attachment":[{"href":"https:\/\/loopbeatzz.info\/wp\/wp-json\/wp\/v2\/media?parent=62237"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/loopbeatzz.info\/wp\/wp-json\/wp\/v2\/categories?post=62237"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/loopbeatzz.info\/wp\/wp-json\/wp\/v2\/tags?post=62237"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}