Feature Extraction and Learning Using Context Cue and Rényi Entropy Based Mutual Information

详细信息    查看全文

• 关键词：Context Kernel Descriptors ; Cauchy ; Schwarz Quadratic Mutual Information ; Feature extraction and learning ; Object classification and detection
• 刊名：Lecture Notes in Computer Science
• 出版年：2015
• 出版时间：2015
• 年：2015
• 卷：9493
• 期：1
• 页码：69-88
• 全文大小：2,766 KB
• 参考文献：1.Bo, L., Ren, X., Fox, D.: Kernel descriptors for visual recognition. In: NIPS, pp. 244–252 (2010)
  2.Bo, L., Lai, K., Ren, X., Fox, D.: Object recognition with hierarchical kernel descriptors. In: CVPR, vol. 1, pp. 1729–1736 (2011)
  3.Bo, L., Sminchisescu, C.: Efficient match kernel between sets of features for visual recognition. In: NIPS, vol. 1, pp. 135–143 (2009)
  4.Wang, P., et al.: Supervised kernel descriptor for visual recognition. In: CVPR, vol. 1, pp. 2858–2865 (2013)
  5.Jégou, H., Douze, M., Schmid, C.: Packing bag-of-features. In: ICCV, vol. 1, pp. 2357–2364 (2009)
  6.Cao, Y. et al.: Spatial-bag-of-features. In: CVPR, vol. 1, pp. 3352–3359 (2010)
  7.Lazebnik, S., Schmid, C., Ponce, J.: Beyond bags of features: Spatial pyramid matching for recognizing natural scene categories. In: CVPR, vol. 1, pp. 2169–2178 (2006)
  8.Lowe, D.: Distinctive image features from scale-invariant keypoints. IJCV 60(2), 91–110 (2004)CrossRef
  9.Bay, H., Ess, A., Tuytelaars, T., Gool, L.: Van.: SURF: speeded up robust features. Comput. Vis. Image Underst. 110(3), 346–359 (2008)CrossRef
  10.Ojala, T., Pietikäinen, M., Mäenpää, T.: Multiresolution gray-scale and rotation invariant texture classification with local binary patterns. IEEE Trans. PAMI 24(7), 971–987 (2002)CrossRef
  11.Dalal, N., Triggs, B.: Histograms of oriented gradients for human detection. In: CVPR, vol. 1, pp. 886–893 (2005)
  12.Pedersen, K., Smidt, K., Ziem, A., Igel, C.: Shape index descriptors applied to texture-based galaxy analysis. In: ICCV, vol. 1, pp. 2240–2447 (2013)
  13.Alcantarilla, P.F., Bartoli, A., Davison, A.J.: KAZE features. In: Fitzgibbon, A., Lazebnik, S., Perona, P., Sato, Y., Schmid, C. (eds.) ECCV 2012, Part VI. LNCS, vol. 7577, pp. 214–227. Springer, Heidelberg (2012)CrossRef
  14.Scholkopf, B., Smola, A., Mulle, K.: Kernel principal component analysis. In: ICANN, vol. 1327, pp. 583–588 (1997)
  15.Scholkopf, B., Smola, A., Mulle, K.: Nonlinear component analysis as a kernel eigenvalue problem. Neural Comput. 10(5), 1299–1319 (1998)CrossRef
  16.Battiti, R.: Using mutual information for selecting features in supervised neural net learning. IEEE Trans. Neural Netw. 5(4), 537–550 (1994)CrossRef
  17.Peng, H., Long, F., Ding, C.: Feature selection based on mutual information: criteria of max-dependency, max-relevance, and min-redundancy. IEEE Trans. PAMI 27(8), 1226–1238 (2005)CrossRef
  18.Yang, H., Moody, J.: Feature selection based on joint mutual information. Int. ICSC Symp. Adv. Intell. Data Anal. vol. 1, pp. 22–25 (1999)
  19.Kwak, N., Choi, C.: Input feature selection by mutual information based on parzen window. IEEE Trans. PAMI 24(12), 1667–1671 (2002)CrossRef
  20.Zhang, Z., Hancock, E.R.: A graph-based approach to feature selection. In: Jiang, X., Ferrer, M., Torsello, A. (eds.) GbRPR 2011. LNCS, vol. 6658, pp. 205–214. Springer, Heidelberg (2011)CrossRef
  21.Liu, C., Shum, H.: Kullback-Leibler boosting. In: CVPR, vol. 1, pp. 587–594 (2003)
  22.Qiu, Q., Patel, V., Chellappa, R.: Information-theoretic dictionary learning for image classification. IEEE Trans. PAMI 36(11), 2173–2184 (2014)CrossRef
  23.Brown, G., Pocock, A., Zhao, M., Luján, M.: Conditional likelihood maximisation: a unifying framework for information theoretic feature selection. J. Mach. Learn. Res. 13(1), 27–66 (2012)MathSciNet MATH
  24.Leiva, J., Artes, A.: Information-theoretic linear feature extraction based on kernel density estimators: a review. IEEE Trans. Syst. Man Cybern. Part C Appl. Rev. 42(6), 1180–1189 (2012)CrossRef
  25.Hild II, K., Erdogmus, D., Principe, J.: An analysis of entropy estimators for blind source separation. Sign. Proces. 86(1), 182–194 (2006)CrossRef MATH
  26.Hild II, K., Erdogmus, D., Torkkola, K., Principe, J.: Feature extraction using information-theoretic learning. IEEE Trans. PAMI 28(9), 1385–1392 (2006)CrossRef
  27.Rényi, A.: On measures of entropy and information. In: Fourth Berkeley Symposium on Mathematical Statistics and Probability, pp. 547–561 (1961)
  28.Principe, J.: Information theoretic learning: Renyi’s entropy and kernel perspectives. Springer, Heidelberg (2010)CrossRef
  29.Parzen, E.: On the estimation of a probability density function and the mode. Ann. Math. Statist. 33(3), 1065–1076 (1962)MathSciNet CrossRef MATH
  30.Jenssen, R.: Kernel entropy component analysis. IEEE Trans. PAMI 32(5), 847–860 (2010)CrossRef
  31.Jenssen, R., Eltoft, T.: A new information theoretic analysis of sum-of-squared-error kernel clustering. Neurocomputing 72(1–3), 23–31 (2008)CrossRef
  32.Gómez, L., Jenssen, R., Camps-Valls, G.: Kernel entropy component analysis for remote sensing image clustering. IEEE Geosci. Remote Sens. Lett. 9(2), 312–316 (2012)CrossRef
  33.Zhong, Z., Hancock, E.: Kernel entropy-based unsupervised spectral feature selection. Int. J. Pattern Recogn. Artif. Intell. 26(5), 126002-1-18 (2012)
  34.Hellman, M., Raviv, J.: Probability of error, equivocation, and the Chernoff bound. IEEE Trans. Inf. Theor. 16(4), 368–372 (1979)MathSciNet CrossRef
  35.Georghiades, A., Belhumeur, P., Kriegman, D.: From few to many: Ilumination cone models for face recognition under variable lighting and pose. IEEE Trans. PAMI 23, 643–660 (2001)CrossRef
  36.Li, F., Fergus, R., Perona, P.: One-shot learning of object categories. IEEE Trans. PAMI 28(4), 594–611 (2006)CrossRef
  37.Torralba, A., Fergus, R., Freeman, W.: 80 million tiny images: A large data set for nonparametric object and scene recognition. IEEE Trans. PAMI 30(11), 1958–1970 (2008)CrossRef
  38.Mika, S., et al.: Fisher discriminant analysis with kernels. In: IEEE Neural Networks for Signal Processing Workshop, pp. 41–48 (1999)
  39.He, X., et al.: Face recognition using laplacianfaces. IEEE Trans. PAMI 27(3), 328–340 (2005)CrossRef
  40.Jia, Y., Huang, C., Darrell, T.: Beyond spatial pyramids: Receptive field learning for pooled image features. In: CVPR, vol. 1, pp. 3370–3377 (2012)
  41.Jiang, Z., Zhang, G., Davis, L.: Submodular dictionary learning for sparse coding. In: CVPR, vol. 1, pp. 3418–3425 (2012)
  42.Boureau, Y., et al.: Ask the locals: Multi-way local pooling for image recognition. In: ICCV, vol. 1, pp. 2651–2658 (2011)
  43.Liu, L., et al.: In defense of soft-assignment coding. In: ICCV, pp. 2486–2493 (2011)
  44.Zeiler, M., Taylor, W., Fergus, R.: Adaptive deconvolutional networks for mid and high level feature learning. In: ICCV, vol. 1, pp. 2018–2025 (2011)
  45.Feng, J., Ni, B., Tian, Q., Yan, S.: Geometric p-norm feature pooling for image classification. In: CVPR, vol. 1, pp. 2697–2704 (2011)
  46.Oliveira, G., Nascimento, E., Vieira, A.: Sparse spatial coding: a novel approach for efficient and accurate object recognition. In: ICRA, pp. 2592–2598 (2012)
  47.McCann, S., Lowe, D.G.: Spatially local coding for object recognition. In: Lee, K.M., Matsushita, Y., Rehg, J.M., Hu, Z. (eds.) ACCV 2012, Part I. LNCS, vol. 7724, pp. 204–217. Springer, Heidelberg (2013)CrossRef
  48.Bo, L., Ren, X., Fox, D.: Multipath sparse coding using hierarchical matching pursuit. In: CVPR, vol. 1, pp. 660–667 (2013)
  49.Seidenari, L., Serra, G., Bagdanov, A., Del Bimbo, A.: Local pyramidal descriptors for image recognition. IEEE Trans. PAMI 36(5), 1033–1040 (2014)CrossRef
  50.Bosch, A., Zisserman, A., Munoz, X.: Image classification using random forests and ferns. In: ICCV, vol. 1, pp. 1–8 (2007)
  51.Ciresan, D., Meier, U., Schmidhuber, J.: Multi-column deep neural networks for image classification. In: CVPR, pp. 3642–3649 (2012)
  52.Zeiler, M., Fergus, R.: Stochastic pooling for regularization of deep convolutional neural networks. In: ICLR (2013)
  53.Le, Q., et al.: Tiled convolutional neural networks. In: NIPS, vol. 1, pp. 1279–1287 (2010)
  54.Yu, K., Zhang, T.: Improved local coordinate coding using local tangents. In: ICML, vol. 1, pp. 1215–1222 (2010)
  55.Goodfellow, I., Courville, A., Bengio, Y.: Spike-and-slab sparse coding for unsupervised feature discovery. In: NIPS Workshop on Challenges in Learning Hierarchical Models (2011)
  56.Everingham, M., et al.: The pascal visual object classes (VOC) challenge. Int. J. Comput. Vis. 88(2), 303–338 (2010)CrossRef
  
• 作者单位：Hong Pan (16) (17)
  Søren Ingvor Olsen (16)
  Yaping Zhu (16)
  
  16. Department of Computer Science, University of Copenhagen, 2100, København Ø, Denmark
  17. School of Automation, Southeast University, Nanjing, 210096, China
  
• 丛书名：Pattern Recognition: Applications and Methods
• ISBN：978-3-319-27677-9
• 刊物类别：Computer Science
• 刊物主题：Artificial Intelligence and Robotics
  Computer Communication Networks
  Software Engineering
  Data Encryption
  Database Management
  Computation by Abstract Devices
  Algorithm Analysis and Problem Complexity
  
• 出版者：Springer Berlin / Heidelberg
• ISSN：1611-3349

文摘

Feature extraction and learning play a critical role for visual perception tasks. We focus on improving the robustness of the kernel descriptors (KDES) by embedding context cues and further learning a compact and discriminative feature codebook for feature reduction using Rényi entropy based mutual information. In particular, for feature extraction, we develop a new set of kernel descriptors−Context Kernel Descriptors (CKD), which enhance the original KDES by embedding the spatial context into the descriptors. Context cues contained in the context kernel enforce some degree of spatial consistency, thus improving the robustness of CKD. For feature learning and reduction, we propose a novel codebook learning method, based on a Rényi quadratic entropy based mutual information measure called Cauchy-Schwarz Quadratic Mutual Information (CSQMI), to learn a compact and discriminative CKD codebook. Projecting the original full-dimensional CKD onto the codebook, we reduce the dimensionality of CKD while preserving its discriminability. Moreover, the latent connection between Rényi quadratic entropy and the mapping data in kernel feature space further facilitates us to capture the geometric structure as well as the information about the underlying labels of the CKD using CSQMI. Thus the resulting codebook and reduced CKD are discriminative. We verify the effectiveness of our method on several public image benchmark datasets such as YaleB, Caltech-101 and CIFAR-10, as well as a challenging chicken feet dataset of our own. Experimental results show that our method has promising potential for visual object recognition and detection applications.