基于多特征结合与支持向量机集成的噪声检测与图像去噪
|
摘要
图像去噪是图像处理中关键的预处理环节。由于在含噪图像中,大部分噪声和图像细节分布在高频区域,不易区分,导致去噪时会不同程度地损坏图像的细节信息。因此,如何能在去噪的同时最大限度地保持细节信息是图像去噪研究的重点。
针对在仅依赖单个图像特征时,基于支持向量机(Support Vector Machine, SVM)的图像去噪方法未能获得较好的去噪效果且会导致噪声点的识别率较低、分类器性能较差的问题,本文在分析总结图像去噪相关算法的基础上,探索了一种基于多特征结合与支持向量机的图像去噪方法。该方法根据图像中相邻像素的相关性及椒盐噪声的特点,利用多种特征相结合的方式来全面的描述像素点的属性,从而准确地区分噪声点和非噪声点。实验结果表明,基于多特征结合与支持向量机的图像去噪方法的去噪效果更优。此外,相对于基于单个特征的支持向量机分类器,基于多特征结合的支持向量机分类器对噪声点的识别率较高,且分类器性能更优。
鉴于支持向量机集成(Support Vector Machine Ensemble, SVM Ensemble)的分类性能优于单个支持向量机,且稳定性、泛化能力更好,本文将支持向量机集成应用于图像去噪,提出了一种基于多特征结合与支持向量机集成的图像去噪方法。首先,根据图像中相邻像素的相关性及椒盐噪声的特点,提取含噪图像中的多种特征,并将其结合构成样本集;其次,对样本集进行归一化处理,并采用同时扰动训练样本和分类器模型参数的二重扰动机制及多数投票法构造支持向量机集成分类器;然后,利用支持向量机集成分类器识别含噪图像中的噪声点,再利用支持向量回归机对噪声点的灰度值进行回归预测;最后,重构图像达到去噪的目的。实验结果表明,该方法能进一步提高去噪效果,且在去噪的同时能较好地保持图像的细节信息,并在低噪声比下尤为有效。此外,与基于多特征结合的支持向量机分类器相比,基于多特征结合的支持向量机集成分类器具有更好的分类性能、稳定性和泛化能力。
Image de-noising is a key pretreatment link before image is processed. For noisy image, most noise and image details are hard to be distinguished because they distribute in high frequency area, which leads to the loss of image detail information in different degree while noise points are removed. Therefore, how to retain image details to the utmost extent while noise points are removed is a research focus.
Aiming at the problem that the approach for removing noise from images based on SVM fails to gain better de-noising effect and leads to low performance of classifier and noise recognition rate when only rely on a single image feature, an image de-noising method based on multi-feature combination and SVM is proposed on the basis of analyzing and summarizing the algorithms related to image de-noising. According to the adjacent pixels correlation in the image and the characteristics of salt-pepper noise, the method comprehensively describes the pixels by using the way of multi-feature combination. Thus noise points and non-noise points can be differentiated accurately. Experimental results show that the image de-noising method based on multi-feature combination and SVM gains better de-noising effect. In addition, the support vector classifier based on multi-feature combination has higher noise recognition rate and classification performance compared with support vector classifier based on single image feature.
Considering that SVM ensemble has better classification performance, stability and generalization capability, SVM ensemble is applied to image de-noising. A new approach for removing salt-pepper noise based on multi-feature combination and SVM ensemble is presented in this thesis. First, according to the correlation of adjacent pixels in the image and the characteristics of salt-pepper noises, multiple features are extracted from noisy image and constituted sample set. Second, the sample set is normalized. Then the dual disturbance mechanism and majority voting method are applied to construct SVM ensemble classifier. The dual disturbance mechanism disturbs training set and classifier model parameter. Third, the SVM ensemble classifier is used to recognize noise points and support vector regression is applied to predicting the original gray value of noise points. Finally, image is reconstructed so as to achieve the purpose of de-noising. Experimental results show that this approach can further improve de-noising effect, and can retain detail information of image better while noise points are removed. The method is especially effective in lower noise ratio. Moreover, the SVM ensemble classifier based on multi-feature combination has better classification, stability and generalization capability compared with the SVM classifier based on multi-feature combination.
针对在仅依赖单个图像特征时,基于支持向量机(Support Vector Machine, SVM)的图像去噪方法未能获得较好的去噪效果且会导致噪声点的识别率较低、分类器性能较差的问题,本文在分析总结图像去噪相关算法的基础上,探索了一种基于多特征结合与支持向量机的图像去噪方法。该方法根据图像中相邻像素的相关性及椒盐噪声的特点,利用多种特征相结合的方式来全面的描述像素点的属性,从而准确地区分噪声点和非噪声点。实验结果表明,基于多特征结合与支持向量机的图像去噪方法的去噪效果更优。此外,相对于基于单个特征的支持向量机分类器,基于多特征结合的支持向量机分类器对噪声点的识别率较高,且分类器性能更优。
鉴于支持向量机集成(Support Vector Machine Ensemble, SVM Ensemble)的分类性能优于单个支持向量机,且稳定性、泛化能力更好,本文将支持向量机集成应用于图像去噪,提出了一种基于多特征结合与支持向量机集成的图像去噪方法。首先,根据图像中相邻像素的相关性及椒盐噪声的特点,提取含噪图像中的多种特征,并将其结合构成样本集;其次,对样本集进行归一化处理,并采用同时扰动训练样本和分类器模型参数的二重扰动机制及多数投票法构造支持向量机集成分类器;然后,利用支持向量机集成分类器识别含噪图像中的噪声点,再利用支持向量回归机对噪声点的灰度值进行回归预测;最后,重构图像达到去噪的目的。实验结果表明,该方法能进一步提高去噪效果,且在去噪的同时能较好地保持图像的细节信息,并在低噪声比下尤为有效。此外,与基于多特征结合的支持向量机分类器相比,基于多特征结合的支持向量机集成分类器具有更好的分类性能、稳定性和泛化能力。
Image de-noising is a key pretreatment link before image is processed. For noisy image, most noise and image details are hard to be distinguished because they distribute in high frequency area, which leads to the loss of image detail information in different degree while noise points are removed. Therefore, how to retain image details to the utmost extent while noise points are removed is a research focus.
Aiming at the problem that the approach for removing noise from images based on SVM fails to gain better de-noising effect and leads to low performance of classifier and noise recognition rate when only rely on a single image feature, an image de-noising method based on multi-feature combination and SVM is proposed on the basis of analyzing and summarizing the algorithms related to image de-noising. According to the adjacent pixels correlation in the image and the characteristics of salt-pepper noise, the method comprehensively describes the pixels by using the way of multi-feature combination. Thus noise points and non-noise points can be differentiated accurately. Experimental results show that the image de-noising method based on multi-feature combination and SVM gains better de-noising effect. In addition, the support vector classifier based on multi-feature combination has higher noise recognition rate and classification performance compared with support vector classifier based on single image feature.
Considering that SVM ensemble has better classification performance, stability and generalization capability, SVM ensemble is applied to image de-noising. A new approach for removing salt-pepper noise based on multi-feature combination and SVM ensemble is presented in this thesis. First, according to the correlation of adjacent pixels in the image and the characteristics of salt-pepper noises, multiple features are extracted from noisy image and constituted sample set. Second, the sample set is normalized. Then the dual disturbance mechanism and majority voting method are applied to construct SVM ensemble classifier. The dual disturbance mechanism disturbs training set and classifier model parameter. Third, the SVM ensemble classifier is used to recognize noise points and support vector regression is applied to predicting the original gray value of noise points. Finally, image is reconstructed so as to achieve the purpose of de-noising. Experimental results show that this approach can further improve de-noising effect, and can retain detail information of image better while noise points are removed. The method is especially effective in lower noise ratio. Moreover, the SVM ensemble classifier based on multi-feature combination has better classification, stability and generalization capability compared with the SVM classifier based on multi-feature combination.
引文
[1]姚敏.数字图像处理[M].北京:机械工业出版社, 2008
[2] Rafael C. Gonzalez, Richard E. Woods.阮秋琦,阮宇智等译.冈萨雷斯数字图像处理(第二版)[M].北京:电子工业出版社, 2005
[3]王晓凯,李锋.改进的自适应中值滤波[J].计算机工程与应用, 2010, 46 (3): 175~176, 218
[4]邢藏菊,王守觉,邓浩江等.一种基于极值中值的新型滤波算法[J].中国图形图像学报, 2001, 6 (6): 533~563
[5] Sun T, Neuvo Y. Detail-preserving Median Based Filters in Image Processing[J]. Pattern Recognit. Lett, 1994, 15 (4): 341~347
[6]赵春晖,乔景渌,孙圣和.一类多结构元自适应广义形态滤波器[J].中国图像图形学报, 1997, 2 (11): 806~810
[7]李玲玲,余文勇.基于数学形态学的灰度线形态识别研究与开发[J].计算机工与应用, 2002, 11: l04~l06
[8]段汕,秦前清.卷积形态滤波与图像去噪[J].计算机工程与应用, 2007, 43 (13): 37~40
[9]于盛林,刘文波.用于减少随机误差的中值—模糊滤波器[J].计量学报, 1995, 16 (4): 297~300
[10]杨群生,陈敏,余英林.基于模糊技术的随机噪声消除算法[J].华南理工大学学报,2000,28 (8):82~87
[11]桂晟.基于模糊技术的混合滤波算法研究与实现[D].北京邮电大学硕士学位论文, 2008
[12]章琳,方志军,汪胜前等.基于遗传算法的多小波自适应去噪方法研究[J].红外与毫米波学报, 2009, 28 (1): 77~80
[13]王志明,伍朝华.基于改进遗传算法的小波去噪的阈值优化[J].计算机工程与设计, 2008, 29 (9): 2381~2383
[14]顾晓东,郭仕德,余道衡.一种基于PCNN的图像去噪新方法[J].电子与信息学报, 2002, 24 (10): 1304~1308
[15]刘远民,秦世引.一种新的基于PCNN的自适应强去噪方法[J].北京航空航天大学学报, 2009, 35 (01): 108~112
[16] Donoho D L, Johnstone I M. Ideal Spatial Adaptation Via Wavelet Shrinkage[J]. Biometrika, 1994, 81 (12): 425~455
[17] Donoho D L, Johnstone I M. Adapting to Unknown Smoothness Via Wavelet Shrinkage[J]. Journal of the American Statistical Association, 1995, 90 (432): 1200~1224
[18] Perona P, Malik J. Scale-space and Edge Detection Using Anisotropic Diffusion[J]. IEEE Transactions on Pattem Analysis and Machine Intelligence, 1990, 12 (7): 629~639
[19] Alvarez L, Lions P L, Morel J M. Image Selective Smoothing and Edge Detection by Nonlinear Diffusion[J]. SIAM Journal on numerical analysis, 1992, 29 (3): 845~866
[20]孙红星,赵楠楠,徐心如.基于SVM的模糊推理在图像降噪中的建模与仿真[J].系统仿真学报, 2006, 18 (11): 3276~3279
[21]杨朝晖,陈映鹰.基于支持向量机的椒盐噪声去除方法[J].计算机工程与应用, 2009, 45 (22): 150~152
[22]冼广铭,曾碧卿.ε-支持向量回归机算法及其应用[J].计算机工程与应用, 2008, 44 (17): 40~42
[23] Cortes C, Vapnik V. Support Vector Networks[J]. Machine Learning, 1995, 20: 273~297
[24] Sch?lkopf B et al. New Support Vector Algorithms[J]. Neural Computation, 2000, 12: 1207~1245
[25] Suykens. J A K, Vandewalle J. Least Squares Support Vector Machine Classifiers[J]. Neural Processing Letters, 1999, 9 (3): 293~300
[26] Roobaert D. DirectSVM: a fast and simple support vector machine perception[C]. Proceedings of the 2000 IEEE signal Processing Society Workshop, 2000, 356~365
[27] Lin ChunFu, Wang ShengDe. Fuzzy Support Vector Machines[J]. IEEE Transaction on Neural Network, 2002, 3 (2): 464~471
[28] Fung G, Mangasarian O L. Proximal Support Vector Machine Classifiers[C]. Proc 7th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2001
[29] K. Veropoulos, C. Campbell, N. Cristianini. Controlling the Sensitivity of Support Vector Machines[C]. Proceedings of the International Joint Conference on AI, 1999: 55~60
[30]刘莉,陈秀宏.改进的支持向量机分类算法[J].计算机工程与设计, 2009, 30 (11): 2763~2765
[31] Thomas G Dietterich. Ensemble Methods in Machine Learning[J]. Lecture Notes in Computer Science, Calgari, Italy, 2000, 1857: 1~15
[32] Kim H C, Pang S N, Je H M et al. Constructing Support Vector Machine Ensemble[J]. Pattern Recognition, 2003, 36 (12): 2757~2767
[33] C. A. M. Lima, A. L. V. Coelho, F. J. Von Zuben. Ensembles of Support Vector Machines for Regression Problems[C]. Procs. INNS-IEEE International Joint Conference on Neural Networks, 2002, 3: 2381~2386
[34] B. Y. Sun, D. S. Huang. Least Squares Support Vector Machines Ensemble[C]. International Joint Conference on Neural Network, 2004: 2013~2016
[35] LingMing He, XiaoBing Yang, HuiJun Lu. A Comparison of Support Vector Machines Ensemble for Classification[C]. Proceedings of the Sixth International Conference on Machine Learning and Cybernetics, 2007: 3613~3617
[36] Dimitrios S, Fossyniotis, Andreas Stafylopatis. A Multi-SVM Classification System[J]. LNCS, 2001, 2096: 198~207
[37] G.Valentini. An Experimental Bias-Variance Anylysis of SVM Ensembles Based on Resampling Techniques[J]. IEEE transactions on System, Man and Cybernetics-Part B: Cybernetics, 2005, 35 (6): 1252~1271
[38] Xu Chun Li, Lei Wang, Eric Sung. AdaBoost with SVM Based Component Classifiers[J]. Egnineering Application of Artificial Intelligence. 2008 (21): 785~795
[39]何灵敏.支持向量机集成及其在遥感分类中的应用[D].浙江大学博士学位论文, 2006
[40]李国正,李丹.集成学习中特征选择技术[J].上海大学学报(自然科学版), 2007, 13 (5): 598~604
[41]何鸣,李国正,袁捷等.基于主成分分析的Bagging集成学习方法[J].上海大学学报(自然科学版), 2006, 12 (4): 415~418, 427
[42]贾华丁,游志胜,王磊.采用二重扰动机制的支持向量机的集成训练算法[J].控制与决策, 2008, 23 (7): 828~832
[43]蔡铁,伍星,李烨.集成学习中基于离散化方法的基分类器构造研究[J].计算机应用, 2008, 28 (8): 2091~2093
[44]边肇祺,张学工等.模式识别(第二版)[M].北京:清华大学出版社, 2007
[45]邓乃扬,田英杰.数据挖掘中的新方法——支持向量机[M].北京:科学出版社, 2006
[46] L. Breiman. Bagging Predictors[J]. Machine Learning, 1996, 24 (3): 123~140
[47]杨高波,杜青松. MATLAB图像/视频处理应用及实例[M].北京:电子工业出版社, 2010
[48]朱嘉钢. SVR的鲁棒性及其在图像恢复中的应用研究[D].南京理工大学博士学位论文, 2005
[49]杨海军,梁德群,江学峰.基于方向信息测度的自适应多级中值滤波器[J].电子与信息学报, 2001, 23 (12): 1326~1332
[50]李双全.基于中值滤波和小波变换的图像去噪算法研究[D].哈尔滨理工大学博士学位论文, 2009
[51] Chih-Chung Chang, Chih-Jen Lin. LIBSVM: a library for support vector machines[EB/OL]. http://www.csie.ntu.edu.tw/~cjlin/libsvm, 2001
[2] Rafael C. Gonzalez, Richard E. Woods.阮秋琦,阮宇智等译.冈萨雷斯数字图像处理(第二版)[M].北京:电子工业出版社, 2005
[3]王晓凯,李锋.改进的自适应中值滤波[J].计算机工程与应用, 2010, 46 (3): 175~176, 218
[4]邢藏菊,王守觉,邓浩江等.一种基于极值中值的新型滤波算法[J].中国图形图像学报, 2001, 6 (6): 533~563
[5] Sun T, Neuvo Y. Detail-preserving Median Based Filters in Image Processing[J]. Pattern Recognit. Lett, 1994, 15 (4): 341~347
[6]赵春晖,乔景渌,孙圣和.一类多结构元自适应广义形态滤波器[J].中国图像图形学报, 1997, 2 (11): 806~810
[7]李玲玲,余文勇.基于数学形态学的灰度线形态识别研究与开发[J].计算机工与应用, 2002, 11: l04~l06
[8]段汕,秦前清.卷积形态滤波与图像去噪[J].计算机工程与应用, 2007, 43 (13): 37~40
[9]于盛林,刘文波.用于减少随机误差的中值—模糊滤波器[J].计量学报, 1995, 16 (4): 297~300
[10]杨群生,陈敏,余英林.基于模糊技术的随机噪声消除算法[J].华南理工大学学报,2000,28 (8):82~87
[11]桂晟.基于模糊技术的混合滤波算法研究与实现[D].北京邮电大学硕士学位论文, 2008
[12]章琳,方志军,汪胜前等.基于遗传算法的多小波自适应去噪方法研究[J].红外与毫米波学报, 2009, 28 (1): 77~80
[13]王志明,伍朝华.基于改进遗传算法的小波去噪的阈值优化[J].计算机工程与设计, 2008, 29 (9): 2381~2383
[14]顾晓东,郭仕德,余道衡.一种基于PCNN的图像去噪新方法[J].电子与信息学报, 2002, 24 (10): 1304~1308
[15]刘远民,秦世引.一种新的基于PCNN的自适应强去噪方法[J].北京航空航天大学学报, 2009, 35 (01): 108~112
[16] Donoho D L, Johnstone I M. Ideal Spatial Adaptation Via Wavelet Shrinkage[J]. Biometrika, 1994, 81 (12): 425~455
[17] Donoho D L, Johnstone I M. Adapting to Unknown Smoothness Via Wavelet Shrinkage[J]. Journal of the American Statistical Association, 1995, 90 (432): 1200~1224
[18] Perona P, Malik J. Scale-space and Edge Detection Using Anisotropic Diffusion[J]. IEEE Transactions on Pattem Analysis and Machine Intelligence, 1990, 12 (7): 629~639
[19] Alvarez L, Lions P L, Morel J M. Image Selective Smoothing and Edge Detection by Nonlinear Diffusion[J]. SIAM Journal on numerical analysis, 1992, 29 (3): 845~866
[20]孙红星,赵楠楠,徐心如.基于SVM的模糊推理在图像降噪中的建模与仿真[J].系统仿真学报, 2006, 18 (11): 3276~3279
[21]杨朝晖,陈映鹰.基于支持向量机的椒盐噪声去除方法[J].计算机工程与应用, 2009, 45 (22): 150~152
[22]冼广铭,曾碧卿.ε-支持向量回归机算法及其应用[J].计算机工程与应用, 2008, 44 (17): 40~42
[23] Cortes C, Vapnik V. Support Vector Networks[J]. Machine Learning, 1995, 20: 273~297
[24] Sch?lkopf B et al. New Support Vector Algorithms[J]. Neural Computation, 2000, 12: 1207~1245
[25] Suykens. J A K, Vandewalle J. Least Squares Support Vector Machine Classifiers[J]. Neural Processing Letters, 1999, 9 (3): 293~300
[26] Roobaert D. DirectSVM: a fast and simple support vector machine perception[C]. Proceedings of the 2000 IEEE signal Processing Society Workshop, 2000, 356~365
[27] Lin ChunFu, Wang ShengDe. Fuzzy Support Vector Machines[J]. IEEE Transaction on Neural Network, 2002, 3 (2): 464~471
[28] Fung G, Mangasarian O L. Proximal Support Vector Machine Classifiers[C]. Proc 7th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2001
[29] K. Veropoulos, C. Campbell, N. Cristianini. Controlling the Sensitivity of Support Vector Machines[C]. Proceedings of the International Joint Conference on AI, 1999: 55~60
[30]刘莉,陈秀宏.改进的支持向量机分类算法[J].计算机工程与设计, 2009, 30 (11): 2763~2765
[31] Thomas G Dietterich. Ensemble Methods in Machine Learning[J]. Lecture Notes in Computer Science, Calgari, Italy, 2000, 1857: 1~15
[32] Kim H C, Pang S N, Je H M et al. Constructing Support Vector Machine Ensemble[J]. Pattern Recognition, 2003, 36 (12): 2757~2767
[33] C. A. M. Lima, A. L. V. Coelho, F. J. Von Zuben. Ensembles of Support Vector Machines for Regression Problems[C]. Procs. INNS-IEEE International Joint Conference on Neural Networks, 2002, 3: 2381~2386
[34] B. Y. Sun, D. S. Huang. Least Squares Support Vector Machines Ensemble[C]. International Joint Conference on Neural Network, 2004: 2013~2016
[35] LingMing He, XiaoBing Yang, HuiJun Lu. A Comparison of Support Vector Machines Ensemble for Classification[C]. Proceedings of the Sixth International Conference on Machine Learning and Cybernetics, 2007: 3613~3617
[36] Dimitrios S, Fossyniotis, Andreas Stafylopatis. A Multi-SVM Classification System[J]. LNCS, 2001, 2096: 198~207
[37] G.Valentini. An Experimental Bias-Variance Anylysis of SVM Ensembles Based on Resampling Techniques[J]. IEEE transactions on System, Man and Cybernetics-Part B: Cybernetics, 2005, 35 (6): 1252~1271
[38] Xu Chun Li, Lei Wang, Eric Sung. AdaBoost with SVM Based Component Classifiers[J]. Egnineering Application of Artificial Intelligence. 2008 (21): 785~795
[39]何灵敏.支持向量机集成及其在遥感分类中的应用[D].浙江大学博士学位论文, 2006
[40]李国正,李丹.集成学习中特征选择技术[J].上海大学学报(自然科学版), 2007, 13 (5): 598~604
[41]何鸣,李国正,袁捷等.基于主成分分析的Bagging集成学习方法[J].上海大学学报(自然科学版), 2006, 12 (4): 415~418, 427
[42]贾华丁,游志胜,王磊.采用二重扰动机制的支持向量机的集成训练算法[J].控制与决策, 2008, 23 (7): 828~832
[43]蔡铁,伍星,李烨.集成学习中基于离散化方法的基分类器构造研究[J].计算机应用, 2008, 28 (8): 2091~2093
[44]边肇祺,张学工等.模式识别(第二版)[M].北京:清华大学出版社, 2007
[45]邓乃扬,田英杰.数据挖掘中的新方法——支持向量机[M].北京:科学出版社, 2006
[46] L. Breiman. Bagging Predictors[J]. Machine Learning, 1996, 24 (3): 123~140
[47]杨高波,杜青松. MATLAB图像/视频处理应用及实例[M].北京:电子工业出版社, 2010
[48]朱嘉钢. SVR的鲁棒性及其在图像恢复中的应用研究[D].南京理工大学博士学位论文, 2005
[49]杨海军,梁德群,江学峰.基于方向信息测度的自适应多级中值滤波器[J].电子与信息学报, 2001, 23 (12): 1326~1332
[50]李双全.基于中值滤波和小波变换的图像去噪算法研究[D].哈尔滨理工大学博士学位论文, 2009
[51] Chih-Chung Chang, Chih-Jen Lin. LIBSVM: a library for support vector machines[EB/OL]. http://www.csie.ntu.edu.tw/~cjlin/libsvm, 2001