基于图结构的图像分割算法研究
|
摘要
图像分割是计算机视觉的最底层功能,占有很重要的地位。而且,现有的图像分割的方法都没有良好的普适性,因此,研究一种良好的图像分割方法,对计算机视觉来说是极为重要的。
图像分割方法按其分割的模式可以分为有监督的图像分割与无监督的图像分割。有监督的图像分割由于是基于人机的交互,也称交互式的图像分割。这种分割含有人的先验知识,其分割速度快,而且效果也要好,其分割的难度也相对较低。近几年来,随着技术的不断成熟,交互式图像分割的效果显著提高,已经基本达到分割的理想效果,并且在图像编辑等领域已经有了很广泛的应用。而无监督图像分割是完全依靠算法的能力对图像实现自动分割。由于没有人工的介入,该方式完全靠算法来分割出图像中的物体,其难度比较大。目前分割方法主要有确定型方法和概率型方法两大类。确定型方法速度快,但精确度往往较低,对精细复杂的图像的分割容易丢失信息;概率型方法则是有较高的精确度,对原图像有很准确的分割效果,但速度较慢。
本文采用Potts模型的图结构模型,提出了基于概率机制的改进型算法。经典的Potts模型功能强大,但将其用于图像分割时最显著的缺点是模型十分耗时,收敛速度相当慢,主要原因在于Potts模型存在临界状态慢转移过程,Ralf Opera提出了采用Potts模型的基于能量的聚类更新算法,一定程度上弥补了Potts模型的收敛慢的缺点。本文的主要工作是针对图像分割的特点,在采用能量聚类更新的基础上,首先采用了图像的预分割,借鉴Swendsen-Wang算法的特长,将对像素的处理转换为对原子像素团的处理,从而大大加快了聚类的过程;然后改进了算法实现的采样方法,用Metropolis采样替代原来的Gibbs采样,从而加速了模型的收敛。另外,为了进一步提高算法的普适性,本文对模型的能量函数作了修改,先利用Gabor滤波提取图像的纹理信息,并作为能量函数的一部分,从而改善了对纹理图像的分割效果。最后,本文在静态图像的分割的基础上,对视频图像序列作了分割,并得到了预期的分割效果。
Image segmentation is the basic feature of computer vision. As there are no existing segmentation algorithms suitable for all the types of images, it is extremely important to develop algorithms that are robust and universally suited.
There are two types of segmentation methods: supervised and unsupervised. The supervised method completes the segmentation course with some human interaction. It is also called the interactive image segmentation. This method usually achieves better segmentation results because of the prior information given by the user. So far, this type of segmentation has been well developed and become more and more mature, and it has been used in different areas such as image editing. Another type of segmentation method is called the unsupervised segmentation. This method is to segment the image automatically with no human interaction. It belongs to the low-level vision and is the basic feature of the vision system. With this part well designed, the high-level vision will be well done with semantics explanation. This type of method is much more difficult than the supervised one does. There are two directions following this method, the deterministic method and the probabilistic method. The deterministic method has higher speed and lower precision while comparing with the probabilistic method.
The algorithm proposed in this paper is developed using the probabilistic mechanics based on the Potts model. The classical Potts model is a powerful tool for image segmentation but the drawback of the model is its slow convergence. The main reason for this is that there exists a critical slowing down process at phase transitions. To overcome the drawback of the Potts model, an image segmentation method based on the ECU (Energy based Cluster Update) algorithm is designed according to the characteristics of image segmentation. Firstly, with the preprocessing of merging single pixels into atomic regions, the image is further segmented with these atomic regions in stead of pixels, thus greatly accelerates the whole segmentation process. Secondly, the Metropolis sampler is adopted to speed up the sampling and the convergence of the model. Finally, the algorithm is successfully applied to segment both the static images and video sequence images. Experimental results show that the proposed method is robust and quite applicable.
图像分割方法按其分割的模式可以分为有监督的图像分割与无监督的图像分割。有监督的图像分割由于是基于人机的交互,也称交互式的图像分割。这种分割含有人的先验知识,其分割速度快,而且效果也要好,其分割的难度也相对较低。近几年来,随着技术的不断成熟,交互式图像分割的效果显著提高,已经基本达到分割的理想效果,并且在图像编辑等领域已经有了很广泛的应用。而无监督图像分割是完全依靠算法的能力对图像实现自动分割。由于没有人工的介入,该方式完全靠算法来分割出图像中的物体,其难度比较大。目前分割方法主要有确定型方法和概率型方法两大类。确定型方法速度快,但精确度往往较低,对精细复杂的图像的分割容易丢失信息;概率型方法则是有较高的精确度,对原图像有很准确的分割效果,但速度较慢。
本文采用Potts模型的图结构模型,提出了基于概率机制的改进型算法。经典的Potts模型功能强大,但将其用于图像分割时最显著的缺点是模型十分耗时,收敛速度相当慢,主要原因在于Potts模型存在临界状态慢转移过程,Ralf Opera提出了采用Potts模型的基于能量的聚类更新算法,一定程度上弥补了Potts模型的收敛慢的缺点。本文的主要工作是针对图像分割的特点,在采用能量聚类更新的基础上,首先采用了图像的预分割,借鉴Swendsen-Wang算法的特长,将对像素的处理转换为对原子像素团的处理,从而大大加快了聚类的过程;然后改进了算法实现的采样方法,用Metropolis采样替代原来的Gibbs采样,从而加速了模型的收敛。另外,为了进一步提高算法的普适性,本文对模型的能量函数作了修改,先利用Gabor滤波提取图像的纹理信息,并作为能量函数的一部分,从而改善了对纹理图像的分割效果。最后,本文在静态图像的分割的基础上,对视频图像序列作了分割,并得到了预期的分割效果。
Image segmentation is the basic feature of computer vision. As there are no existing segmentation algorithms suitable for all the types of images, it is extremely important to develop algorithms that are robust and universally suited.
There are two types of segmentation methods: supervised and unsupervised. The supervised method completes the segmentation course with some human interaction. It is also called the interactive image segmentation. This method usually achieves better segmentation results because of the prior information given by the user. So far, this type of segmentation has been well developed and become more and more mature, and it has been used in different areas such as image editing. Another type of segmentation method is called the unsupervised segmentation. This method is to segment the image automatically with no human interaction. It belongs to the low-level vision and is the basic feature of the vision system. With this part well designed, the high-level vision will be well done with semantics explanation. This type of method is much more difficult than the supervised one does. There are two directions following this method, the deterministic method and the probabilistic method. The deterministic method has higher speed and lower precision while comparing with the probabilistic method.
The algorithm proposed in this paper is developed using the probabilistic mechanics based on the Potts model. The classical Potts model is a powerful tool for image segmentation but the drawback of the model is its slow convergence. The main reason for this is that there exists a critical slowing down process at phase transitions. To overcome the drawback of the Potts model, an image segmentation method based on the ECU (Energy based Cluster Update) algorithm is designed according to the characteristics of image segmentation. Firstly, with the preprocessing of merging single pixels into atomic regions, the image is further segmented with these atomic regions in stead of pixels, thus greatly accelerates the whole segmentation process. Secondly, the Metropolis sampler is adopted to speed up the sampling and the convergence of the model. Finally, the algorithm is successfully applied to segment both the static images and video sequence images. Experimental results show that the proposed method is robust and quite applicable.
引文
[1] Gonzalez .Rafael C.Digital Image Processing. 2nd ed.北京:电子工业出版社,2007.
[2]刘中合.数字图像处理技术现状与展望.计算机时代, 2005,9:6-8.
[3]李红俊.数字图像处理技术及其应用.计算机测量与控制,2002,9: 620-622.
[4]罗希平.图像分割方法综述.模式识别与人工智能,1999,12(3): 300-312.
[5] Nikhil R. A review on image segmentation techniques.Pattern Recognition, 1993, 26(9): 1277-1294.
[6] Zhang Y.J.A survey on Evaluation methods for Image Segmentation.Pattern Recognition, 1996, 29(8): 1335-1346.
[7] Sahoo .P.K.A Survey of Thresholding Techniques.Computer Vision, Graphics and Image Processing, 1995, 41:233-260.
[8] Yen, Jui-cheng.A New Criteria For Automatic Multilevel Thresholding.IEEE trans on Image Processing, 1995, 4(3):370-377.
[9] Pikas.A . Digital Image Thresholding Based on Topological Stable-state . Pattern Recognition, 1996, 29(5):829-843.
[10] Papamarkos N.A New Approach for Multilevel Threshold Selection.GVGIP: Models and Image Processing, 1994, 56:357-370.
[11] Huang L.K.Image Thresholding by Minimizing the Measure of Fuzziness.Pattern Recognition, 1995, 28(1): 41-51.
[12] Corneloup.G.Bscan Image segmentation by Thresholding Using Coccurrence Matrix analysis.Pattern Recognition, 1996, 29(2): 281-296.
[13] Li .L.Grey-level Image Thresholding Based on Fisher Linear Projection of two dimensional Histogram.Pattern Recognition, 1997,30(5):743-750.
[14] Sahoo.P . Threshold Selection Using Renyi’s Entropy . Pattern Recognition, 1997,30(1):71-84.
[15] Brink A D.Thresholding of Digital Images Using Two Dimensional Entropy.Pattern Recognition,1993, 25(8):803-808.
[16] Cheng H.D.Threshold Selection Based on Fuzzy c-partition Entropy Approach.Pattern Recognition, 1998, 31(7):857-870.
[17] Haralick R.M . Digital Step Edges From Zero Crossing of Second Directional Derivatives.IEEE Trans on Pattern Analysis and Machine Intelligence, 1984,6(1):58-68.
[18] Dorin Comaniciu. Mean Shift: A Robust Approach Toward Feature Space Analysis. IEEE TRANSECTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, MAY 2002,VOL.24, NO.5.
[19] Hong Y.P . Improved Mean Shift Segmentation Approach for Natural Images.International Conference on Intelligent Computing, 2005, 185(13): 940-952.
[20] Wang Jue.Image and Video Segmentation by Anisotropic Kernel Mean Shift.In European Conf. on Computer Vision, 2004.
[21] Collins.R.Mean Shift blob Tracking Through Scale Space.In IEEE Conf. on Computer Vision and Pattern Recognition, 2004.
[22] Cheng. Y.Mean Shift, Mode Seeking, and Clustering.IEEE Trans on Pattern Analysis and Machine Intelligence, 1995, 17:790-799.
[23] Comaniciu. D.An Algorithm for Data-Driven Bandwidth Selection.IEEE Trans on Pattern Analysis and Machine Intelligence, 2003, 25(2):281-288.
[24] Comaniciu.D.Kernel Based Object Tracking.IEEE Trans on Pattern Analysis and Machine Intelligence, 2003, 25:564-575.
[25] Matalas L.An Edge Detection Techniques Using the Facet Model and Parameterized Relaxation Labeling.IEEE Trans on Pattern Analysis and Machine Intelligence,1997, 19(4): 328-341.
[26] Canny J.A Computational Approach to Edge Detection.IEEE Trans on Pattern Analysis and Machine Intelligence, 1986, 8(6):679-698.
[27] Gary Chartand.Introduction to Graph Theory.北京:人民邮电出版社,2007.
[28]徐俊明.图论及其应用.合肥:中国科技大学出版社,1998.
[29]闫成新.基于图论的图像分割研究进展.计算机工程与应用,2006,5: 11-14.
[30] Yuri Boykov.An Experimental Comparison of Min Cut/Max Flow Algorithms for Energy Minimization in Computer Vison.In IEEE Trans on PAMI, 2004, 26(9):1124-1137.
[31] Yuri Boykov.A New Bayesian Framework for Object Recognition.In IEEE Conference on Computer Vision and Pattern Recognition, 1999, 2: 517-523.
[32] Yuri Boykov . Computing Geodesics and Minimal Surface via Graph Cut . In International Conference on Computer Vision, 2003, 1: 26-33.
[33] Yuri. Boykov. Interactive Graph Cuts for Optimal Boundary & Region Segmentation of Objects in N-D Images.International Conference on Computer Vision, Vancouver, Canada, July 2001, 1: 105.
[34] Yuri. Boykov.Fast Approximate Energy Minimization via Graph Cuts.Proceedings of the seventh IEEE International Conference on Computer Vision, 1999, 1:377-384.
[35] Zayed.N.M . Wavelet Segmentation for Fetal Ultrasound Images . MWSCAS , Proceedings of the 44th IEEE 2001 Midwest Symposium on, 2001, 1:501-504.
[36] Hideki.Noda . MRF Based Texture Segmentation Using Wavelet decomposed Images.Pattern Recognition, 2002, 35(4):771-782.
[37] Bouman C.A.Multi resolution Segmentation of Textured Images.IEEE Trans on Pattern Analysis and Machine Intelligence, 1991, 13(2):99-113.
[38] Bouman C.A . A Multi Scale Random Field Model for Bayesian Image Segmentation.IEEE Trans on Image Process, 1994, 3(2):162-177.
[39] Cormer.M.L . Segmentation of Texture Image Using a Multiresolution Gaussian Autoregressive Model.IEEE Trans on Image Process, 1999, 8(3):408-418.
[40]江涛、朱光喜等,基于小波变换的点量化图像编码算法,计算机辅助设计与图形学学报, 2002,14(10):897-904。
[41] Salari E.Texture Segmentation Using Hierarchical Wavelet Decomposition.Pattern Recognition,1995, 28(12):1819-1824.
[42] Komodakis. N.A New Framework for Approximate Labeling via Graph Cuts.ICCV, Beijing, 2005, 1018-1025.
[43] Pearl. J.Fusion, propagation, and structuring in belief networks.Artif. Intell, 1986, pp. 241-288.
[44] V. Vapnik.The Nature of Statistical Learning Theory.Springer-Verlag, 1995.
[45] Potts. R.Some Generalized Order-Disorder Transformations.Proceedings of the Cambridge Philosophical Society, 1952, Vol.48, pp.106-109.
[46] F.Y.Wu.The Potts Model.Reviews of Modern Physics, January 1982, vol.54,no.1.
[47] Laura Beaudin. A Review of the Potts Model. Ros-Hulman Institute of Technology math journal 2007, vol8-n1.
[48] Opera. R.A Fast and Robust Cluster Update Algorithm for Image Segmentation in Spin-lattice Models Without Annealing–Visual Lattice revisited.Nueral Computation, August 1998, vol.10, pp.6.
[49] Ferber .C.v.Cluster Update and recognition.Physical Review E - Statistical Physics, Plasmas, Fluids, and Related Interdisciplinary Topics, 2000, 62(2 B), pp.1461-1464.
[50] Jain. A.K.Unsupervised Texture Segmentation Using Gabor Filters.Pattern recognition, 1991, 24(12): 1167-1186.
[51] Jordi Freixenet . Colour Texture Segmentation by Region- Boundary Cooperation.ECCV, 2004, 2: 250-261.
[52] Weldon. T.P . Efficient Gabor Filter Design for Texture Segmentation . Pattern Recognition, 1996, 29(12): 2005-2015.
[53] Andriun Christophe.An Introduction to MCMC for Machine Learning. Machine Learning, 2003, 50: 5-43.
[54] Radford M.Neal . Probabilistic Inference Using Markov Chain Monte Carlo Methods.1993.
[55] Ashkin J.Statistics of Two-Dimensional Lattices with Four Components.Physical Review. 1943, Vol.64, No.5 and 6.
[56] Adrian Barbu.Graph Partition by Swendsen-Wang Cuts.Computer Vision, 2003. Proceedings. Ninth IEEE International Conference on Volume , Issue , 13-16 Oct. 2003 vol.1, Page(s): 320 - 327.
[57] Barbu. A . Generalizing Swendsen-Wang to Sampling Arbitrary Posterior Probabilities.IEEE Transactions on Pattern Analysis and Machine Intelligence, August 2005, 27( 8): 1239 - 1253.
[58] Adrian Barbu.On the relationship between image and motion segmentation. Computer Science, 3667 LNCS, pp. 51-63.
[59] George S.Fishman . An Analysis of Swendsen-Wang and Related Sampling Methods.J.R Statist Soc.B(1999) 61,Part 3,pp.623-641.
[60] Mark Huber .A Bounding Chain for Swendsen-Wang .Random Structures and Algorithms, 22(1): 43-59.
[61] Tu .Z.W.Parsing Images into Regions, Curves, and Curve Groups.Int'l Journal of Computer Vision, August, 2006, 69(2): 223-249.
[62] Tu Zhuowen.An Integrated Framework for Image Segmentation and Perceptual Grouping.10th IEEE International Conference on Computer Vision(ICCV),Oct,2005.
[63] Tu Zhuowen.Image Segmentation by Data Driven Markov Chain Monte Carlo.8th IEEE International Conference on Computer Vision(ICCV),July,2001.
[64] Zhu Song-Chun.Integrating Bottom-Up/Top-Down for Object Recognition by Data Driven Markov Chain Monte Carlo.IEEE Computer Vision and Pattern Recognition(CVPR), July,2000.
[65] Swendsen. R. H.Replica Monte Carlo Simulation of Spin-glasses.Phys. Rev. Lett, 1986 {57}, pp.2607.
[66] Swendsen. R. H.Nonuniversal Critical Dynamics in Monte Carlo Simulations.Phys. Rev. Lett, 1987, 58: pp.86.
[67] Wang. J.-S.Monte Carlo Renormalization-group Study of Ising Spin Glasses.Phys. Rev. B ,1988, 37: 7745 .
[68] Wang . J.-S.Antiferromagnetic Potts Models.Phys. Rev. Lett. 1989, 63: 109 .
[69] Jianbo Shi.Normalized Cuts and Image Segmentation.IEEE TRANSECTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE , AUGUST 2000, VOL 22, NO 8,.
[70] Wang Yang.Spatiotemporal Video Segmentation Based on Graphical Models. IEEE TRANSACTIONS ON IMAGE PROCESSING, July2005, VOL. 14, NO. 7.
[71] Pei Yin.Tree Based Classifier for Bilayer Video Segmentation. IEEE Computer Vision and Pattern Recognition(CVPR),July, 2007
[72] Deng Y.N .Unsupervised Segmentation of Color-texture Regions in Images and Video,.In IEEE Trans. on Pattern Analysis and Machine Intelligence, 2001,23(8):800-810.
[73] Yilmaz. A.Contour Based Object Tracking With Occlusion Handling in Video Aquired Using Mobile Cameras.IEEE Trans on Pattern Analysis and Machine Intelligence, 2004, 26(11):1531-1536.
[74] Zhao Jieyu.Video Object Segmentation with a Potts Model.Proceedings of the Third International Conference on Natural Computation, 2007, 2:742-748.
[75] W. Phillips III.Flame Recognition in Video.Pattern Recognition Letters,Elsevier, 2002, 23 (3):319-327.
[76] Rosenfeld.A.Scene Labeling by Relaxation Operation.IEEE Trans on System, man and Cybernetics, 6:420-433.
[77] Waltz D.M . Generating Semantic Descriptions From Drawing of Scenes with Shadows.The Psychology of Computer Vision, 1972, New York, 19-92.
[78] Eric.C.Rouchka.A Brief Overview of Gibbs Sampling.Washington University Institute for Biomedical Computing Statistics Study Group, May, 20, 1997.
[79] Tang Yinggan .Multi-resolution Image Segmentation based on Gaussian Mixture Model.Journal of System and Engineering and Electronics, 2006, 17(4):870-874.
[80] Zhao Jieyu . Segmenting Moving Objects with a Recurrent Stochastic Neural Network . Proceedings of the 11th International Conference on Neural Information Processing(ICONIP2004), Lecture Notes in Computer Science, Springer-Verlag, 2004.
[2]刘中合.数字图像处理技术现状与展望.计算机时代, 2005,9:6-8.
[3]李红俊.数字图像处理技术及其应用.计算机测量与控制,2002,9: 620-622.
[4]罗希平.图像分割方法综述.模式识别与人工智能,1999,12(3): 300-312.
[5] Nikhil R. A review on image segmentation techniques.Pattern Recognition, 1993, 26(9): 1277-1294.
[6] Zhang Y.J.A survey on Evaluation methods for Image Segmentation.Pattern Recognition, 1996, 29(8): 1335-1346.
[7] Sahoo .P.K.A Survey of Thresholding Techniques.Computer Vision, Graphics and Image Processing, 1995, 41:233-260.
[8] Yen, Jui-cheng.A New Criteria For Automatic Multilevel Thresholding.IEEE trans on Image Processing, 1995, 4(3):370-377.
[9] Pikas.A . Digital Image Thresholding Based on Topological Stable-state . Pattern Recognition, 1996, 29(5):829-843.
[10] Papamarkos N.A New Approach for Multilevel Threshold Selection.GVGIP: Models and Image Processing, 1994, 56:357-370.
[11] Huang L.K.Image Thresholding by Minimizing the Measure of Fuzziness.Pattern Recognition, 1995, 28(1): 41-51.
[12] Corneloup.G.Bscan Image segmentation by Thresholding Using Coccurrence Matrix analysis.Pattern Recognition, 1996, 29(2): 281-296.
[13] Li .L.Grey-level Image Thresholding Based on Fisher Linear Projection of two dimensional Histogram.Pattern Recognition, 1997,30(5):743-750.
[14] Sahoo.P . Threshold Selection Using Renyi’s Entropy . Pattern Recognition, 1997,30(1):71-84.
[15] Brink A D.Thresholding of Digital Images Using Two Dimensional Entropy.Pattern Recognition,1993, 25(8):803-808.
[16] Cheng H.D.Threshold Selection Based on Fuzzy c-partition Entropy Approach.Pattern Recognition, 1998, 31(7):857-870.
[17] Haralick R.M . Digital Step Edges From Zero Crossing of Second Directional Derivatives.IEEE Trans on Pattern Analysis and Machine Intelligence, 1984,6(1):58-68.
[18] Dorin Comaniciu. Mean Shift: A Robust Approach Toward Feature Space Analysis. IEEE TRANSECTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, MAY 2002,VOL.24, NO.5.
[19] Hong Y.P . Improved Mean Shift Segmentation Approach for Natural Images.International Conference on Intelligent Computing, 2005, 185(13): 940-952.
[20] Wang Jue.Image and Video Segmentation by Anisotropic Kernel Mean Shift.In European Conf. on Computer Vision, 2004.
[21] Collins.R.Mean Shift blob Tracking Through Scale Space.In IEEE Conf. on Computer Vision and Pattern Recognition, 2004.
[22] Cheng. Y.Mean Shift, Mode Seeking, and Clustering.IEEE Trans on Pattern Analysis and Machine Intelligence, 1995, 17:790-799.
[23] Comaniciu. D.An Algorithm for Data-Driven Bandwidth Selection.IEEE Trans on Pattern Analysis and Machine Intelligence, 2003, 25(2):281-288.
[24] Comaniciu.D.Kernel Based Object Tracking.IEEE Trans on Pattern Analysis and Machine Intelligence, 2003, 25:564-575.
[25] Matalas L.An Edge Detection Techniques Using the Facet Model and Parameterized Relaxation Labeling.IEEE Trans on Pattern Analysis and Machine Intelligence,1997, 19(4): 328-341.
[26] Canny J.A Computational Approach to Edge Detection.IEEE Trans on Pattern Analysis and Machine Intelligence, 1986, 8(6):679-698.
[27] Gary Chartand.Introduction to Graph Theory.北京:人民邮电出版社,2007.
[28]徐俊明.图论及其应用.合肥:中国科技大学出版社,1998.
[29]闫成新.基于图论的图像分割研究进展.计算机工程与应用,2006,5: 11-14.
[30] Yuri Boykov.An Experimental Comparison of Min Cut/Max Flow Algorithms for Energy Minimization in Computer Vison.In IEEE Trans on PAMI, 2004, 26(9):1124-1137.
[31] Yuri Boykov.A New Bayesian Framework for Object Recognition.In IEEE Conference on Computer Vision and Pattern Recognition, 1999, 2: 517-523.
[32] Yuri Boykov . Computing Geodesics and Minimal Surface via Graph Cut . In International Conference on Computer Vision, 2003, 1: 26-33.
[33] Yuri. Boykov. Interactive Graph Cuts for Optimal Boundary & Region Segmentation of Objects in N-D Images.International Conference on Computer Vision, Vancouver, Canada, July 2001, 1: 105.
[34] Yuri. Boykov.Fast Approximate Energy Minimization via Graph Cuts.Proceedings of the seventh IEEE International Conference on Computer Vision, 1999, 1:377-384.
[35] Zayed.N.M . Wavelet Segmentation for Fetal Ultrasound Images . MWSCAS , Proceedings of the 44th IEEE 2001 Midwest Symposium on, 2001, 1:501-504.
[36] Hideki.Noda . MRF Based Texture Segmentation Using Wavelet decomposed Images.Pattern Recognition, 2002, 35(4):771-782.
[37] Bouman C.A.Multi resolution Segmentation of Textured Images.IEEE Trans on Pattern Analysis and Machine Intelligence, 1991, 13(2):99-113.
[38] Bouman C.A . A Multi Scale Random Field Model for Bayesian Image Segmentation.IEEE Trans on Image Process, 1994, 3(2):162-177.
[39] Cormer.M.L . Segmentation of Texture Image Using a Multiresolution Gaussian Autoregressive Model.IEEE Trans on Image Process, 1999, 8(3):408-418.
[40]江涛、朱光喜等,基于小波变换的点量化图像编码算法,计算机辅助设计与图形学学报, 2002,14(10):897-904。
[41] Salari E.Texture Segmentation Using Hierarchical Wavelet Decomposition.Pattern Recognition,1995, 28(12):1819-1824.
[42] Komodakis. N.A New Framework for Approximate Labeling via Graph Cuts.ICCV, Beijing, 2005, 1018-1025.
[43] Pearl. J.Fusion, propagation, and structuring in belief networks.Artif. Intell, 1986, pp. 241-288.
[44] V. Vapnik.The Nature of Statistical Learning Theory.Springer-Verlag, 1995.
[45] Potts. R.Some Generalized Order-Disorder Transformations.Proceedings of the Cambridge Philosophical Society, 1952, Vol.48, pp.106-109.
[46] F.Y.Wu.The Potts Model.Reviews of Modern Physics, January 1982, vol.54,no.1.
[47] Laura Beaudin. A Review of the Potts Model. Ros-Hulman Institute of Technology math journal 2007, vol8-n1.
[48] Opera. R.A Fast and Robust Cluster Update Algorithm for Image Segmentation in Spin-lattice Models Without Annealing–Visual Lattice revisited.Nueral Computation, August 1998, vol.10, pp.6.
[49] Ferber .C.v.Cluster Update and recognition.Physical Review E - Statistical Physics, Plasmas, Fluids, and Related Interdisciplinary Topics, 2000, 62(2 B), pp.1461-1464.
[50] Jain. A.K.Unsupervised Texture Segmentation Using Gabor Filters.Pattern recognition, 1991, 24(12): 1167-1186.
[51] Jordi Freixenet . Colour Texture Segmentation by Region- Boundary Cooperation.ECCV, 2004, 2: 250-261.
[52] Weldon. T.P . Efficient Gabor Filter Design for Texture Segmentation . Pattern Recognition, 1996, 29(12): 2005-2015.
[53] Andriun Christophe.An Introduction to MCMC for Machine Learning. Machine Learning, 2003, 50: 5-43.
[54] Radford M.Neal . Probabilistic Inference Using Markov Chain Monte Carlo Methods.1993.
[55] Ashkin J.Statistics of Two-Dimensional Lattices with Four Components.Physical Review. 1943, Vol.64, No.5 and 6.
[56] Adrian Barbu.Graph Partition by Swendsen-Wang Cuts.Computer Vision, 2003. Proceedings. Ninth IEEE International Conference on Volume , Issue , 13-16 Oct. 2003 vol.1, Page(s): 320 - 327.
[57] Barbu. A . Generalizing Swendsen-Wang to Sampling Arbitrary Posterior Probabilities.IEEE Transactions on Pattern Analysis and Machine Intelligence, August 2005, 27( 8): 1239 - 1253.
[58] Adrian Barbu.On the relationship between image and motion segmentation. Computer Science, 3667 LNCS, pp. 51-63.
[59] George S.Fishman . An Analysis of Swendsen-Wang and Related Sampling Methods.J.R Statist Soc.B(1999) 61,Part 3,pp.623-641.
[60] Mark Huber .A Bounding Chain for Swendsen-Wang .Random Structures and Algorithms, 22(1): 43-59.
[61] Tu .Z.W.Parsing Images into Regions, Curves, and Curve Groups.Int'l Journal of Computer Vision, August, 2006, 69(2): 223-249.
[62] Tu Zhuowen.An Integrated Framework for Image Segmentation and Perceptual Grouping.10th IEEE International Conference on Computer Vision(ICCV),Oct,2005.
[63] Tu Zhuowen.Image Segmentation by Data Driven Markov Chain Monte Carlo.8th IEEE International Conference on Computer Vision(ICCV),July,2001.
[64] Zhu Song-Chun.Integrating Bottom-Up/Top-Down for Object Recognition by Data Driven Markov Chain Monte Carlo.IEEE Computer Vision and Pattern Recognition(CVPR), July,2000.
[65] Swendsen. R. H.Replica Monte Carlo Simulation of Spin-glasses.Phys. Rev. Lett, 1986 {57}, pp.2607.
[66] Swendsen. R. H.Nonuniversal Critical Dynamics in Monte Carlo Simulations.Phys. Rev. Lett, 1987, 58: pp.86.
[67] Wang. J.-S.Monte Carlo Renormalization-group Study of Ising Spin Glasses.Phys. Rev. B ,1988, 37: 7745 .
[68] Wang . J.-S.Antiferromagnetic Potts Models.Phys. Rev. Lett. 1989, 63: 109 .
[69] Jianbo Shi.Normalized Cuts and Image Segmentation.IEEE TRANSECTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE , AUGUST 2000, VOL 22, NO 8,.
[70] Wang Yang.Spatiotemporal Video Segmentation Based on Graphical Models. IEEE TRANSACTIONS ON IMAGE PROCESSING, July2005, VOL. 14, NO. 7.
[71] Pei Yin.Tree Based Classifier for Bilayer Video Segmentation. IEEE Computer Vision and Pattern Recognition(CVPR),July, 2007
[72] Deng Y.N .Unsupervised Segmentation of Color-texture Regions in Images and Video,.In IEEE Trans. on Pattern Analysis and Machine Intelligence, 2001,23(8):800-810.
[73] Yilmaz. A.Contour Based Object Tracking With Occlusion Handling in Video Aquired Using Mobile Cameras.IEEE Trans on Pattern Analysis and Machine Intelligence, 2004, 26(11):1531-1536.
[74] Zhao Jieyu.Video Object Segmentation with a Potts Model.Proceedings of the Third International Conference on Natural Computation, 2007, 2:742-748.
[75] W. Phillips III.Flame Recognition in Video.Pattern Recognition Letters,Elsevier, 2002, 23 (3):319-327.
[76] Rosenfeld.A.Scene Labeling by Relaxation Operation.IEEE Trans on System, man and Cybernetics, 6:420-433.
[77] Waltz D.M . Generating Semantic Descriptions From Drawing of Scenes with Shadows.The Psychology of Computer Vision, 1972, New York, 19-92.
[78] Eric.C.Rouchka.A Brief Overview of Gibbs Sampling.Washington University Institute for Biomedical Computing Statistics Study Group, May, 20, 1997.
[79] Tang Yinggan .Multi-resolution Image Segmentation based on Gaussian Mixture Model.Journal of System and Engineering and Electronics, 2006, 17(4):870-874.
[80] Zhao Jieyu . Segmenting Moving Objects with a Recurrent Stochastic Neural Network . Proceedings of the 11th International Conference on Neural Information Processing(ICONIP2004), Lecture Notes in Computer Science, Springer-Verlag, 2004.