基于SIFT特征和SVM的场景分类
|
摘要
场景分类是计算机视觉领域中的一个基础处理过程,在模式识别、机器学习、图像内容理解、图像检索等中扮演重要角色。在特征提取中,经典的SIFT特征因具有对图像的平移、旋转、缩放、甚至对仿射变换保持不变性且具有良好的显着性和鲁棒性而得到广泛应用。而作为分类器之一,SVM以小样本统计机器学习理论为基础,无需依赖于设计者的经验及先验知识,避免了神经网络实现中的经验成分,而且SVM算法最终转化为凸优化问题,具有全局最优性等诸多优点。本文尝试将SIFT特征算法和SVM结合用于场景图像分类识别中。
本文主要内容包括：首先对场景分类、SIFT特征、SVM的研究现状进行了概述,然后重点介绍了提取SIFT特征点和SVM分类算法。然后,基于SIFT特征和非线性SVM提出了场景图像分类系统的理论框架。具体工作包括：场景图像的预处理,即将彩色图像变成灰度图像,同时将图像缩放到相同大小尺寸；提出基于SIFT特征和SVM的场景分类算法,先用SIFT算法得到每个场景中各个图像的SIFT特征点,以向量形式存贮,再以这些特征向量作为SVM的原始输入数据进行训练分类。最后,通过数值实验,一方面比较用不同特征点数目的分类正确率,同时还与传统分类方法进行了效果比较。实验结果表明,本文的基于SIFT特征和SVM场景分类在准确率上高于几种传统的方法。
Scene classification is a foundational process in computer vision, and it plays an important role in pattern recognition, machine learning, image content understanding, image retrieve. Classical SIFT feature is an invariant feature to rotation, translation in the field of feature extraction. SIFT Feature is provided with good remarkable and robust. SVM is one of newly and effective classifiers. SVM is based on little sample Statistical Learning Theory and independent on designer's experience. So it avoids empirical in Artificial Neural Networks. SVM is changed to convex optimize, so it guarantees global optimal. This paper presents SVM scene classification method based on SIFT feature.
The main content of this paper includes:first, introduce the current research situation of scene classification, SIFT feature, SVM. Then, based on SIFT feature and SVM, we proposed the theory framework of scene classification. Specific work includes:Firstly, we preprocess the scene images before classification, such as convert the color images to gray images, and scale the images to the same size; Secondly, we present the scene classification algorithm based on SIFT and SVM, SIFT feature points are extracted, they are vectors, and then train SVM by those vectors, At last, recognition work is completed using SIFT feature points and SVM. Our experiment result shows that our algorithm performs better than the one based on geodesic distance.
本文主要内容包括：首先对场景分类、SIFT特征、SVM的研究现状进行了概述,然后重点介绍了提取SIFT特征点和SVM分类算法。然后,基于SIFT特征和非线性SVM提出了场景图像分类系统的理论框架。具体工作包括：场景图像的预处理,即将彩色图像变成灰度图像,同时将图像缩放到相同大小尺寸；提出基于SIFT特征和SVM的场景分类算法,先用SIFT算法得到每个场景中各个图像的SIFT特征点,以向量形式存贮,再以这些特征向量作为SVM的原始输入数据进行训练分类。最后,通过数值实验,一方面比较用不同特征点数目的分类正确率,同时还与传统分类方法进行了效果比较。实验结果表明,本文的基于SIFT特征和SVM场景分类在准确率上高于几种传统的方法。
Scene classification is a foundational process in computer vision, and it plays an important role in pattern recognition, machine learning, image content understanding, image retrieve. Classical SIFT feature is an invariant feature to rotation, translation in the field of feature extraction. SIFT Feature is provided with good remarkable and robust. SVM is one of newly and effective classifiers. SVM is based on little sample Statistical Learning Theory and independent on designer's experience. So it avoids empirical in Artificial Neural Networks. SVM is changed to convex optimize, so it guarantees global optimal. This paper presents SVM scene classification method based on SIFT feature.
The main content of this paper includes:first, introduce the current research situation of scene classification, SIFT feature, SVM. Then, based on SIFT feature and SVM, we proposed the theory framework of scene classification. Specific work includes:Firstly, we preprocess the scene images before classification, such as convert the color images to gray images, and scale the images to the same size; Secondly, we present the scene classification algorithm based on SIFT and SVM, SIFT feature points are extracted, they are vectors, and then train SVM by those vectors, At last, recognition work is completed using SIFT feature points and SVM. Our experiment result shows that our algorithm performs better than the one based on geodesic distance.
引文
[1]Boutell M, Luo J. Review of the state of the art in semantic scene classification[R]. Technical Report, University of Rochester,2002.
[2]A Delorme, G Richard, M Fabre-Thorpe. Rapid Processing of complex natural scenes: A role for the magnocellular visual pathways? [J]. Neuro computing,1999,663-570.
[3]A Chella, M Frixione and S Gaglio. Understanding dynamic scenes [J]. Artificial Intelligence,2000,123:89-132.
[4]Vailaya, Figueiredo M,Jain A. Content-based hierarchical classification of vacation mages,1999[C]. Proceeding of IEEE International Conference on Multimedia Computing and Systems,1999,1:528-523.
[5]Szunner M, Picard R W. Indoor-outdoor image classifition,1998[C]. Proceeding of IEEE International Workshop on Content-based Access of Image and Video Database, in conjunction with ICCV,1998.
[6]Bosch A, X Munoz and J Freixenet. Segmentation and description of natural outdoor scenes [J]. Image and Vision computing 25:727-740,2007.
[7]Andrew Payne and Sameer Singh. Indoor vs. outdoor scene classification in digital photographs [J]. Pattern Recognition,2005,38:1533-1545.
[8]Monay F, Gatica-Perez D. PLSA-based image auto-annotation:constraining the latent space [J]. ACM Multimedia,2004.
[9]Sudderth E, Torralba A, Freeman W, Will sky A. Describing visual scenes using transformed Dirichlet processes [J].NIPS,2005.
[10]Li F, Perona P. A Bayesian hierarchy model for learning natural scene categories [J].CVPR,2005.
[11]Devendran V, Amitabh, Hemalatha Thiagarajan. Invariant moment to scene categorization using Support Vector Machines [J]. IJSC,2008,3:128-133.
[12]A Bosch, A Zisserman, X Munoz. Scene classification via pLSA,2006[C]. Proceedings of the 9th European Conference on Computer Vision, 517-530,2006.
[13]K Hotta. Scene classification based on multi-resolution orientation histogram of Gabor features,2008[C]. Proceedings of the International Conference on Computer Vision Systems,,2008 5008:291-301.
[14]Shokoufandeh A, Marsic I, and Dickinson J. View-based object recognition using saliency maps[J]. Image and Vision Computing,1999,17:445-460.
[15]Lindeberg, T. Detecting salient blob-like image structures and their scales with a scale-space primal sketch:a method for focus-of-attention [J]. International Journal of Computer Vision,1993,11:283-318.
[16]Lindeberg, T. Scale-space theory:A basic tool for analysing structures at different scales [J]. Journal of Applied Statistics,1994,21:224-270.
[17]Mikolajczyk K and Schmidt C. Indexing based on scale invariant interest points [J].In Proceedings of the 8th International Conference on Computer Vision, Vancouver, Canada,2001,525-531.
[18]Mikolajczyk K, Schmidt C. Scale & affine invariant interest point detectors [J]. International Journal of Computer Vision,2004,60:63-86
[19]Lowe D G. Object recognition from local scale-invariant features[J]. In Proceedings of International Conference on Computer Vision,1999,1150-1157.
[20]Lowe David G. Distinctive image features from scale-invariant key points [J]. IJCV,2004,60:91-110.
[21]Johnson A and Hebert M. Object recognition by matching oriented points [J]. In Proceedings of the Conference on Computer Vision and Pattern Recognition, Puerto Rico, USA,1997,684-689.
[22]Vapnik V N. The Nature of Statistical Learning Theory[M]//New York, Springer,1996.
[23]Campbell C. Introduction t o kernel methods. Radial Basis Function Networks: Design and applications [M]//Berlin, Springer-Verlag,2000:155-192.
[24]Mercer J. Function of positive and negative type and their connection with the theory of interval equations [J]. Philos. Trans. Roy. Soc. London,1909, A209:415:446.
[25]Aronszajn N. Theory of reproducing kernels [J]. Transaction of the American Mathematical Society,1950,686:337-404.
[26]Aizerman M A, Braverman E M,Rozonoer L I. Theoretical foundation of potential function method in pattern recognition learning[J]. Automation and Remote Control,1964,25:821-837.
[27]Marie Szafranski, Yves Grandvalet. Composite kernel learning [J].Machine Learning,2010,79:73-103.
[28]Osuna E, Freund R, Girosi G. Improved training algorithm for support vector machines,1997[C]. Proceedings of 1997 IEEE Workshop on Neural Networks for Signal Processing, Amelia Island,1997,276-285.
[29]B Scholkopf, J C Burges, A J Smola. Fast training of support vector machines using sequential minimal optimization,1998[C]. Advances in Kernel Methods Support Vector Learning, Cambridge, MIT Press,1998.
[30]Mangasarian 0 L, Musicant D R. Successive overerlaxation for support vector machines [J].IEEE Transaction on neural Networks,1999,10:1032-1037.
[31]Chang C, Lin C J. Training v-support vector classifiers:theory and algorithm [J]. Neural Computation,2005,13:2119-2147.
[32]Fung G, Mangasarian O L, Proximal support vector machine classifiers,2001 [C]. Proceeding of International Conference of Knowledge Discovery and Data Mining, 2001,77-86.
[33]Johan A K S,Tony V G,Jos D B. Least Squares Support Vector Machines[J]. World Scientific,2002.
[34]Weston J, Watkins C. Multi-class support vector machines[R]. CSD-TR-98-04, London, University of London,1998.
[35]Scholkopf B,Burges C J C, Vapnik V N. Extracting support data for a given task,1995[C]. Proceeding of 1st International Conference on Knowledge Discovery &Data mining, Menlo Park, AAAI Press,1995,252-257.
[36]Blanz V, Scholkopf B. Comparison of view-based object recognition algorithms using realistic 3D models[J].Artificial Neural Networks, Spring-Verlag,1996,251-256.
[37]Bredensteiner E J, Bennett K P. Multicategory classification by support vector machines [J]. Computational Optimizations and APPLICATIONS,1999,53-57.
[38]Burges C J. A tutorial on support vector machines for pattern recognition [J].Data Mining and Knowledge Discovery,1998,2:121-167.
[39]Scholkopf B, Smola A J. New support vector algorithms. Neural Computation,2000, 12:1207-1245.
[40]Smola A J, Scholkopf B. A tutorial on support vector regression[R]. NC-TR, London:University of London,1998.
[41]Smola A J. Regression estimation with support vector learning machines[R]. Munchen, Technische Universidad Munchen,1996.
[42]Drucker H, Burges C J, Kaufman L. Support vector machines regression [J].Advances in Neural Information Processing Systems,1997,9:155-161.
[43]Vapnik V N. Three remarks on the support vector method of function estimation, 1999[C].Advances in kernel Methods:Support Vector Learning, Cambridge, MIT Press,1999,25-42.
[44]Scholkopf B, Williamson R C. Support Vector method for novelty detection [J]. Advances in Neural Information Processing Systems,2000,12:582-588.
[45]Campbell C, Bennett K P. A linear programming approach to novelty detection [J]. Advances in Neural Information Processing Systems,2001,14:395-401.
[46]Vapnik V N, Mukherjee S. Support Vector method for multivariate estimation [J]. Advances in Neural Information Processing Systems,1999,659-665.
[47]Weston J, Grammerman A. Support vector density estimation,1999[C]. Advances in kernel Methods:Support Vector Learning, Cambridge, MIT Press,1999,293-306.
[48]13 Scene categories database. http://vision.cs. princeton. edu/Databsets/SceneClassl3. rar.
[49]Brown M, Lowe D G. Invariant features from inertest point groups,2002[C]. In British Machine Vision Conference,Carddiff, Wales,2002,656-665.
[50]Duabechies I. Ten Lectures on Wavelets [M]//CBMSNSF Regional Conference Series in Appl.Math.61, Philadelphia, SIAM,1992.
[2]A Delorme, G Richard, M Fabre-Thorpe. Rapid Processing of complex natural scenes: A role for the magnocellular visual pathways? [J]. Neuro computing,1999,663-570.
[3]A Chella, M Frixione and S Gaglio. Understanding dynamic scenes [J]. Artificial Intelligence,2000,123:89-132.
[4]Vailaya, Figueiredo M,Jain A. Content-based hierarchical classification of vacation mages,1999[C]. Proceeding of IEEE International Conference on Multimedia Computing and Systems,1999,1:528-523.
[5]Szunner M, Picard R W. Indoor-outdoor image classifition,1998[C]. Proceeding of IEEE International Workshop on Content-based Access of Image and Video Database, in conjunction with ICCV,1998.
[6]Bosch A, X Munoz and J Freixenet. Segmentation and description of natural outdoor scenes [J]. Image and Vision computing 25:727-740,2007.
[7]Andrew Payne and Sameer Singh. Indoor vs. outdoor scene classification in digital photographs [J]. Pattern Recognition,2005,38:1533-1545.
[8]Monay F, Gatica-Perez D. PLSA-based image auto-annotation:constraining the latent space [J]. ACM Multimedia,2004.
[9]Sudderth E, Torralba A, Freeman W, Will sky A. Describing visual scenes using transformed Dirichlet processes [J].NIPS,2005.
[10]Li F, Perona P. A Bayesian hierarchy model for learning natural scene categories [J].CVPR,2005.
[11]Devendran V, Amitabh, Hemalatha Thiagarajan. Invariant moment to scene categorization using Support Vector Machines [J]. IJSC,2008,3:128-133.
[12]A Bosch, A Zisserman, X Munoz. Scene classification via pLSA,2006[C]. Proceedings of the 9th European Conference on Computer Vision, 517-530,2006.
[13]K Hotta. Scene classification based on multi-resolution orientation histogram of Gabor features,2008[C]. Proceedings of the International Conference on Computer Vision Systems,,2008 5008:291-301.
[14]Shokoufandeh A, Marsic I, and Dickinson J. View-based object recognition using saliency maps[J]. Image and Vision Computing,1999,17:445-460.
[15]Lindeberg, T. Detecting salient blob-like image structures and their scales with a scale-space primal sketch:a method for focus-of-attention [J]. International Journal of Computer Vision,1993,11:283-318.
[16]Lindeberg, T. Scale-space theory:A basic tool for analysing structures at different scales [J]. Journal of Applied Statistics,1994,21:224-270.
[17]Mikolajczyk K and Schmidt C. Indexing based on scale invariant interest points [J].In Proceedings of the 8th International Conference on Computer Vision, Vancouver, Canada,2001,525-531.
[18]Mikolajczyk K, Schmidt C. Scale & affine invariant interest point detectors [J]. International Journal of Computer Vision,2004,60:63-86
[19]Lowe D G. Object recognition from local scale-invariant features[J]. In Proceedings of International Conference on Computer Vision,1999,1150-1157.
[20]Lowe David G. Distinctive image features from scale-invariant key points [J]. IJCV,2004,60:91-110.
[21]Johnson A and Hebert M. Object recognition by matching oriented points [J]. In Proceedings of the Conference on Computer Vision and Pattern Recognition, Puerto Rico, USA,1997,684-689.
[22]Vapnik V N. The Nature of Statistical Learning Theory[M]//New York, Springer,1996.
[23]Campbell C. Introduction t o kernel methods. Radial Basis Function Networks: Design and applications [M]//Berlin, Springer-Verlag,2000:155-192.
[24]Mercer J. Function of positive and negative type and their connection with the theory of interval equations [J]. Philos. Trans. Roy. Soc. London,1909, A209:415:446.
[25]Aronszajn N. Theory of reproducing kernels [J]. Transaction of the American Mathematical Society,1950,686:337-404.
[26]Aizerman M A, Braverman E M,Rozonoer L I. Theoretical foundation of potential function method in pattern recognition learning[J]. Automation and Remote Control,1964,25:821-837.
[27]Marie Szafranski, Yves Grandvalet. Composite kernel learning [J].Machine Learning,2010,79:73-103.
[28]Osuna E, Freund R, Girosi G. Improved training algorithm for support vector machines,1997[C]. Proceedings of 1997 IEEE Workshop on Neural Networks for Signal Processing, Amelia Island,1997,276-285.
[29]B Scholkopf, J C Burges, A J Smola. Fast training of support vector machines using sequential minimal optimization,1998[C]. Advances in Kernel Methods Support Vector Learning, Cambridge, MIT Press,1998.
[30]Mangasarian 0 L, Musicant D R. Successive overerlaxation for support vector machines [J].IEEE Transaction on neural Networks,1999,10:1032-1037.
[31]Chang C, Lin C J. Training v-support vector classifiers:theory and algorithm [J]. Neural Computation,2005,13:2119-2147.
[32]Fung G, Mangasarian O L, Proximal support vector machine classifiers,2001 [C]. Proceeding of International Conference of Knowledge Discovery and Data Mining, 2001,77-86.
[33]Johan A K S,Tony V G,Jos D B. Least Squares Support Vector Machines[J]. World Scientific,2002.
[34]Weston J, Watkins C. Multi-class support vector machines[R]. CSD-TR-98-04, London, University of London,1998.
[35]Scholkopf B,Burges C J C, Vapnik V N. Extracting support data for a given task,1995[C]. Proceeding of 1st International Conference on Knowledge Discovery &Data mining, Menlo Park, AAAI Press,1995,252-257.
[36]Blanz V, Scholkopf B. Comparison of view-based object recognition algorithms using realistic 3D models[J].Artificial Neural Networks, Spring-Verlag,1996,251-256.
[37]Bredensteiner E J, Bennett K P. Multicategory classification by support vector machines [J]. Computational Optimizations and APPLICATIONS,1999,53-57.
[38]Burges C J. A tutorial on support vector machines for pattern recognition [J].Data Mining and Knowledge Discovery,1998,2:121-167.
[39]Scholkopf B, Smola A J. New support vector algorithms. Neural Computation,2000, 12:1207-1245.
[40]Smola A J, Scholkopf B. A tutorial on support vector regression[R]. NC-TR, London:University of London,1998.
[41]Smola A J. Regression estimation with support vector learning machines[R]. Munchen, Technische Universidad Munchen,1996.
[42]Drucker H, Burges C J, Kaufman L. Support vector machines regression [J].Advances in Neural Information Processing Systems,1997,9:155-161.
[43]Vapnik V N. Three remarks on the support vector method of function estimation, 1999[C].Advances in kernel Methods:Support Vector Learning, Cambridge, MIT Press,1999,25-42.
[44]Scholkopf B, Williamson R C. Support Vector method for novelty detection [J]. Advances in Neural Information Processing Systems,2000,12:582-588.
[45]Campbell C, Bennett K P. A linear programming approach to novelty detection [J]. Advances in Neural Information Processing Systems,2001,14:395-401.
[46]Vapnik V N, Mukherjee S. Support Vector method for multivariate estimation [J]. Advances in Neural Information Processing Systems,1999,659-665.
[47]Weston J, Grammerman A. Support vector density estimation,1999[C]. Advances in kernel Methods:Support Vector Learning, Cambridge, MIT Press,1999,293-306.
[48]13 Scene categories database. http://vision.cs. princeton. edu/Databsets/SceneClassl3. rar.
[49]Brown M, Lowe D G. Invariant features from inertest point groups,2002[C]. In British Machine Vision Conference,Carddiff, Wales,2002,656-665.
[50]Duabechies I. Ten Lectures on Wavelets [M]//CBMSNSF Regional Conference Series in Appl.Math.61, Philadelphia, SIAM,1992.