融合SVM和快速均值漂移的图像分割算法
|
摘要
图像分割(Image Segmentation)是图像处理研究领域的一个重要问题,是图像分析、特征提取、模式识别等的基础和关键步骤.快速均值漂移图像分割算法(FMS,Fast Mean Shift)虽然能获得良好的分割结果,但是在处理过程中需要人为设定带宽值,而带宽的设定会对分割结果造成一定程度的影响.针对这个问题,提出一种融合支持向量机(Support Vector Machine,SVM)和FMS的图像分割算法(FMS-SVM),先用FMS模型预分割图像,然后用SVM分类器对像素分类,以改善带宽设定产生的影响.实验结果表明,提出的基于SVM和FMS的图像分割算法可以获得良好的分割结果.
Image segmentation is a challenging problem in the domain of image processing,and it is the foundation and critical step of image analysis,feature extraction,pattern recognition,etc. Although fast mean shift( FMS) image segmentation algorithm can obtain the good segmentation results,the segmentation process needs setting bandwidth value,which will affect the segmentation result to a certain degree. This paper proposes an image segmentation algorithm based on Support Vector Machine( SVM) and FMS,namely FMSSVM. First,the FMS model is used for pre-segment of an image,and then SVMclassifier is applied to classify all pixels. In so doing,the effect of iteration bandwidth on segmentation result can be decreased. The Experimental results showthat the proposed FMS-SVMmethod can obtain good segmentation results.
Image segmentation is a challenging problem in the domain of image processing,and it is the foundation and critical step of image analysis,feature extraction,pattern recognition,etc. Although fast mean shift( FMS) image segmentation algorithm can obtain the good segmentation results,the segmentation process needs setting bandwidth value,which will affect the segmentation result to a certain degree. This paper proposes an image segmentation algorithm based on Support Vector Machine( SVM) and FMS,namely FMSSVM. First,the FMS model is used for pre-segment of an image,and then SVMclassifier is applied to classify all pixels. In so doing,the effect of iteration bandwidth on segmentation result can be decreased. The Experimental results showthat the proposed FMS-SVMmethod can obtain good segmentation results.
引文
[1]Zhang Yu-jin.Chinese image engineering[J].Journal of Image and Graphics,1995,1(1):78-83.
[2]Zhang Yu-jin.Image analysis(III)[M].Beijing:Tsinghua University Press,2012:1-27.
[3]Selvathi D,Arulmurgan A,Thamarai Seivi S,et al.MRI image segmentation using unsupervised clustering techniques[C].Sixth International Conference on Computational Intelligence and Multimedia Applications(ICCIMA'05),IEEE,2005:105-110.
[4]Jung Y K,Ho Y S.Traffic parameter extraction using video-based vehicle tracking[C].Intelligent Transportation Systems,Proceedings,1999 IEEE/IEEJ/JSAI International Conference on,IEEE,1999:764-769.
[5]Lemarechal C,Fjortoft R,Marthon P,et al.SAR image segmentation by morphological methods[C].Remote Sensing,International Society for Optics and Photonics(SPIE),1998:111-121.
[6]Zhou De-long,Pan Quan,Zhang Hong-cai,et al.Maximum entropy thresholding algorithm[J].Journal of Software,2001,12(9):1420-1422.
[7]Long Jian-wu,Shen Xuan-jing,Chen Hai-peng.Adaptive minimum error thresholding algorithm[J].Acta Automatica Sinica,2012,38(7):1134-1144.
[8]Lu Jian-feng,Lin Hai,Pan Zhi-geng.Adaptive region growing algorithm in medical images segmentation[J].Journal of ComputerAided Design&Computer Graphics,2005,17(10):2168-2173.
[9]Sun Yang,Luo Yu,Zhou Chang-le,et al.A method based on splitcombining algorithm for the segmentation of the image of tongue[J].Journal of Image and Graphics,2003,8(12):1395-1399.
[10]Qu Bin,Zheng Chang-qiong.A fast boundary tracking method based on the rule of region-rrowth[J].Journal of Sichuan University(Engineering Science Edition),2002,34(2):100-103.
[11]Liu Chen,Zhang Dong.Analysis of edge detection operators and application in medical image[J].Computer Technology and Development,2006,16(8):128-130.
[12]Masooleh MG,Moosavi S A S.An improved fuzzy algorithm forimage segmentation[J].World Academy of Science,Engineering and Technology,2008,38:400-404.
[13]Arivazhagan S,Ganesan L.Texture segmentation using wavelet transform[J].Pattern Recognition Letters,2003,24(16):3197-3203.
[14]Dong G,Xie M.Color clustering and learning for image segmentation based on neural networks[J].IEEE Transactions on Neural Networks,2005,16(4):925-936.
[15]Bai X,Wang W.Saliency-SVM:an automatic approach for image segmentation[J].Neurocomputing,2014,136:243-255.
[16]Zhao Sheng-nan,Wang Wen-jian.Fast mean shift of image segmentation[J].Data Acquisition and Processing,2015,30(1):192-201.
[1]章毓晋.中国图像工程[J].中国图象图形学报,1995,1(1):78-83.
[2]章毓晋.图像分析(第3版)[M].北京,清华大学出版社,2012:1-27.
[6]周德龙,潘泉,张洪才,等.最大熵阈值处理算法[J].软件学报,2001,12(9):1420-1422.
[7]龙建武,申铉京,陈海鹏.自适应最小误差阈值分割算法[J].自动化学报,2012,38(7):1134-1144.
[8]陆剑锋,林海,潘志庚.自适应区域生长算法在医学图像分割中的应用[J].计算机辅助设计与图形学学报,2005,17(10):2168-2173.
[9]孙炀,罗瑜,周昌乐,等.一种基于分裂-合并方法的中医舌像区域分割算法及其实现[J].中国图象图形学报:A辑,2003,8(12):1395-1399.
[10]屈彬,郑昌琼.一种基于区域生长规则的快速边缘跟踪算法[J].四川大学学报:工程科学版,2002,34(2):100-103.
[11]刘晨,张东.边缘检测算子研究及其在医学图像中的应用[J].计算机技术与发展,2006,16(8):128-130.
[16]赵胜男,王文剑.一种快速均值漂移图像分割算法[J].数据采集与处理,2015,30(1):192-201.
[2]Zhang Yu-jin.Image analysis(III)[M].Beijing:Tsinghua University Press,2012:1-27.
[3]Selvathi D,Arulmurgan A,Thamarai Seivi S,et al.MRI image segmentation using unsupervised clustering techniques[C].Sixth International Conference on Computational Intelligence and Multimedia Applications(ICCIMA'05),IEEE,2005:105-110.
[4]Jung Y K,Ho Y S.Traffic parameter extraction using video-based vehicle tracking[C].Intelligent Transportation Systems,Proceedings,1999 IEEE/IEEJ/JSAI International Conference on,IEEE,1999:764-769.
[5]Lemarechal C,Fjortoft R,Marthon P,et al.SAR image segmentation by morphological methods[C].Remote Sensing,International Society for Optics and Photonics(SPIE),1998:111-121.
[6]Zhou De-long,Pan Quan,Zhang Hong-cai,et al.Maximum entropy thresholding algorithm[J].Journal of Software,2001,12(9):1420-1422.
[7]Long Jian-wu,Shen Xuan-jing,Chen Hai-peng.Adaptive minimum error thresholding algorithm[J].Acta Automatica Sinica,2012,38(7):1134-1144.
[8]Lu Jian-feng,Lin Hai,Pan Zhi-geng.Adaptive region growing algorithm in medical images segmentation[J].Journal of ComputerAided Design&Computer Graphics,2005,17(10):2168-2173.
[9]Sun Yang,Luo Yu,Zhou Chang-le,et al.A method based on splitcombining algorithm for the segmentation of the image of tongue[J].Journal of Image and Graphics,2003,8(12):1395-1399.
[10]Qu Bin,Zheng Chang-qiong.A fast boundary tracking method based on the rule of region-rrowth[J].Journal of Sichuan University(Engineering Science Edition),2002,34(2):100-103.
[11]Liu Chen,Zhang Dong.Analysis of edge detection operators and application in medical image[J].Computer Technology and Development,2006,16(8):128-130.
[12]Masooleh MG,Moosavi S A S.An improved fuzzy algorithm forimage segmentation[J].World Academy of Science,Engineering and Technology,2008,38:400-404.
[13]Arivazhagan S,Ganesan L.Texture segmentation using wavelet transform[J].Pattern Recognition Letters,2003,24(16):3197-3203.
[14]Dong G,Xie M.Color clustering and learning for image segmentation based on neural networks[J].IEEE Transactions on Neural Networks,2005,16(4):925-936.
[15]Bai X,Wang W.Saliency-SVM:an automatic approach for image segmentation[J].Neurocomputing,2014,136:243-255.
[16]Zhao Sheng-nan,Wang Wen-jian.Fast mean shift of image segmentation[J].Data Acquisition and Processing,2015,30(1):192-201.
[1]章毓晋.中国图像工程[J].中国图象图形学报,1995,1(1):78-83.
[2]章毓晋.图像分析(第3版)[M].北京,清华大学出版社,2012:1-27.
[6]周德龙,潘泉,张洪才,等.最大熵阈值处理算法[J].软件学报,2001,12(9):1420-1422.
[7]龙建武,申铉京,陈海鹏.自适应最小误差阈值分割算法[J].自动化学报,2012,38(7):1134-1144.
[8]陆剑锋,林海,潘志庚.自适应区域生长算法在医学图像分割中的应用[J].计算机辅助设计与图形学学报,2005,17(10):2168-2173.
[9]孙炀,罗瑜,周昌乐,等.一种基于分裂-合并方法的中医舌像区域分割算法及其实现[J].中国图象图形学报:A辑,2003,8(12):1395-1399.
[10]屈彬,郑昌琼.一种基于区域生长规则的快速边缘跟踪算法[J].四川大学学报:工程科学版,2002,34(2):100-103.
[11]刘晨,张东.边缘检测算子研究及其在医学图像中的应用[J].计算机技术与发展,2006,16(8):128-130.
[16]赵胜男,王文剑.一种快速均值漂移图像分割算法[J].数据采集与处理,2015,30(1):192-201.