Beyond “project and sign” for cosine estimation with binary codes

Many nearest neighbor search algorithms rely on encoding real vectors into binary vectors. The most common strategy projects the vectors onto random directions and takes the sign to produce so-called sketches. This paper discusses the sub-optimality of this choice, and proposes a better encoding strategy based on the quantization and reconstruction points of view. Our second contribution is a novel asymmetric estimator for the cosine similarity. Similar to previous asymmetric schemes, the query is not quantized and the similarity is computed in the compressed domain. Both our contribution leads to improve the quality of nearest neighbor search with binary codes. Its efficiency compares favorably against a recent encoding technique.

[1]  Trevor Darrell,et al.  Learning to Hash with Binary Reconstructive Embeddings , 2009, NIPS.

[2]  Jean-Jacques Fuchs,et al.  Spread representations , 2011, 2011 Conference Record of the Forty Fifth Asilomar Conference on Signals, Systems and Computers (ASILOMAR).

[3]  Florent Perronnin,et al.  Large-scale image retrieval with compressed Fisher vectors , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[4]  Antonio Torralba,et al.  Small codes and large image databases for recognition , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.

[5]  Cordelia Schmid,et al.  Hamming Embedding and Weak Geometric Consistency for Large Scale Image Search , 2008, ECCV.

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

[7]  Petros Boufounos,et al.  Universal Rate-Efficient Scalar Quantization , 2010, IEEE Transactions on Information Theory.

[8]  Matthijs Douze,et al.  Searching in one billion vectors: Re-rank with source coding , 2011, 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[9]  Patrick Gros,et al.  Asymmetric hamming embedding: taking the best of our bits for large scale image search , 2011, ACM Multimedia.

[10]  Andrew Zisserman,et al.  Learning Local Feature Descriptors Using Convex Optimisation , 2014, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[11]  Svetlana Lazebnik,et al.  Iterative quantization: A procrustean approach to learning binary codes , 2011, CVPR 2011.

[12]  Olivier Buisson,et al.  Random maximum margin hashing , 2011, CVPR 2011.

[13]  Cordelia Schmid,et al.  Aggregating local descriptors into a compact image representation , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[14]  Kai Li,et al.  Asymmetric distance estimation with sketches for similarity search in high-dimensional spaces , 2008, SIGIR '08.

[15]  G LoweDavid,et al.  Distinctive Image Features from Scale-Invariant Keypoints , 2004 .

[16]  Hervé Jégou,et al.  Anti-sparse coding for approximate nearest neighbor search , 2011, 2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[17]  Vivek K. Goyal,et al.  Quantized Overcomplete Expansions in IRN: Analysis, Synthesis, and Algorithms , 1998, IEEE Trans. Inf. Theory.

[18]  Cordelia Schmid,et al.  Product Quantization for Nearest Neighbor Search , 2011, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[19]  Moses Charikar,et al.  Similarity estimation techniques from rounding algorithms , 2002, STOC '02.

[20]  Svetlana Lazebnik,et al.  Locality-sensitive binary codes from shift-invariant kernels , 2009, NIPS.

[21]  Antonio Torralba,et al.  Spectral Hashing , 2008, NIPS.

[22]  Hervé Jégou,et al.  Asymmetric Hamming Embedding , 2011, MM 2011.