Inverted Index Compression for Scalable Image Matching

To perform fast image matching against large databases, a Vocabulary Tree (VT) uses an inverted index that maps from each tree node to database images which have visited that node. The inverted index can require gigabytes of memory, which significantly slows down the database server. In this paper, we design, develop, and compare techniques for inverted index compression for image-based retrieval. We show that these techniques significantly reduce memory usage, by as much as 5x, without loss in recognition accuracy. Our work includes fast decoding methods, an offline database reordering scheme that exploits the similarity between images for additional memory savings, and a generalized coding scheme for soft-binned feature descriptor histograms. We also show that reduced index memory permits memory-intensive image matching techniques that boost recognition accuracy.

[1]  Richard Szeliski,et al.  City-Scale Location Recognition , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.

[2]  Bernd Girod,et al.  Robust image retrieval using multiview scalable vocabulary trees , 2009, Electronic Imaging.

[3]  JUSTIN ZOBEL,et al.  Inverted files for text search engines , 2006, CSUR.

[4]  Kenneth Ward Church,et al.  Poisson mixtures , 1995, Natural Language Engineering.

[5]  Cordelia Schmid,et al.  Packing bag-of-features , 2009, 2009 IEEE 12th International Conference on Computer Vision.

[6]  Alistair Moffat,et al.  Inverted Index Compression Using Word-Aligned Binary Codes , 2004, Information Retrieval.

[7]  David Nistér,et al.  Scalable Recognition with a Vocabulary Tree , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).

[8]  Christopher Hunt,et al.  Notes on the OpenSURF Library , 2009 .

[9]  Amir Said,et al.  Comparative analysis of arithmetic coding computational complexity , 2004, Data Compression Conference, 2004. Proceedings. DCC 2004.

[10]  Michael Isard,et al.  Lost in quantization: Improving particular object retrieval in large scale image databases , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.

[11]  Alistair Moffat,et al.  Exploiting clustering in inverted file compression , 1996, Proceedings of Data Compression Conference - DCC '96.

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

[13]  Jean Jyh-Jiun Shann,et al.  Fast query evaluation through document identifier assignment for inverted file-based information retrieval systems , 2006, Inf. Process. Manag..

[14]  Alistair Moffat,et al.  Binary codes for non-uniform sources , 2005, Data Compression Conference.

[15]  Guy E. Blelloch,et al.  Index compression through document reordering , 2002, Proceedings DCC 2002. Data Compression Conference.

[16]  A. Gelbukh,et al.  Compression of Boolean inverted files by document ordering , 2003, International Conference on Natural Language Processing and Knowledge Engineering, 2003. Proceedings. 2003.

[17]  Cordelia Schmid,et al.  A contextual dissimilarity measure for accurate and efficient image search , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.

[18]  Bernd Girod,et al.  Outdoors augmented reality on mobile phone using loxel-based visual feature organization , 2008, MIR '08.