Revisiting Winner Take All (WTA) Hashing for Sparse Datasets

WTA (Winner Take All) hashing has been successfully applied in many large scale vision applications. This hashing scheme was tailored to take advantage of the comparative reasoning (or order based information), which showed significant accuracy improvements. In this paper, we identify a subtle issue with WTA, which grows with the sparsity of the datasets. This issue limits the discriminative power of WTA. We then propose a solution for this problem based on the idea of Densification which provably fixes the issue. Our experiments show that Densified WTA Hashing outperforms Vanilla WTA both in image classification and retrieval tasks consistently and significantly.

[1]  Benjamin Recht,et al.  Random Features for Large-Scale Kernel Machines , 2007, NIPS.

[2]  Anxiao Jiang,et al.  Floating Codes for Joint Information Storage in Write Asymmetric Memories , 2007, 2007 IEEE International Symposium on Information Theory.

[3]  Anshumali Shrivastava,et al.  Probabilistic Hashing Techniques for Big Data , 2015 .

[4]  Kristen Grauman,et al.  Relative attributes , 2011, 2011 International Conference on Computer Vision.

[5]  Jay Yagnik,et al.  The power of comparative reasoning , 2011, 2011 International Conference on Computer Vision.

[6]  A. Yuille,et al.  Dense Scale Invariant Descriptors for Images and Surfaces , 2012 .

[7]  Piotr Indyk,et al.  Approximate nearest neighbors: towards removing the curse of dimensionality , 1998, STOC '98.

[8]  Jianguo Zhang,et al.  The PASCAL Visual Object Classes Challenge , 2006 .

[9]  Jonathon Shlens,et al.  Fast, Accurate Detection of 100,000 Object Classes on a Single Machine , 2013, 2013 IEEE Conference on Computer Vision and Pattern Recognition.

[10]  Yihong Gong,et al.  Linear spatial pyramid matching using sparse coding for image classification , 2009, CVPR.

[11]  Hans-Jörg Schek,et al.  A Quantitative Analysis and Performance Study for Similarity-Search Methods in High-Dimensional Spaces , 1998, VLDB.

[12]  Patrick Haffner,et al.  Support vector machines for histogram-based image classification , 1999, IEEE Trans. Neural Networks.

[13]  Ping Li,et al.  Hashing Algorithms for Large-Scale Learning , 2011, NIPS.

[14]  Antonio Torralba,et al.  LabelMe: A Database and Web-Based Tool for Image Annotation , 2008, International Journal of Computer Vision.

[15]  Andrea Vedaldi,et al.  Vlfeat: an open and portable library of computer vision algorithms , 2010, ACM Multimedia.

[16]  Rajat Raina,et al.  Efficient sparse coding algorithms , 2006, NIPS.

[17]  Ping Li,et al.  Improved Densification of One Permutation Hashing , 2014, UAI.

[18]  Ping Li,et al.  Densifying One Permutation Hashing via Rotation for Fast Near Neighbor Search , 2014, ICML.

[19]  Ping Li,et al.  Theory and applications of b-bit minwise hashing , 2011, Commun. ACM.

[20]  David J. Kriegman,et al.  Locally Uniform Comparison Image Descriptor , 2012, NIPS.