Hashing with Local Combinations of Feature Points and Its Application to Camera-Based Document Image Retrieval — Retrieval in 0 . 14 Second from 10 , 000 Pages —

This paper presents a new method of indexing and retrieval of planar objects based on feature points and its application to document image retrieval using cameras. As the indexing method we propose a method based on local combinations of projective invariants calculated from feature points. As the retrieval method we employ a voting technique for efficiency and robustness against erasure of feature points. Experimental results on 10,000 images with 50 queries show that the method is effective (98% accuracy; the remaining query was ranked at the 5th position among 10,000) and efficient (0.14 second per query).

[1]  D. Forsyth,et al.  Using Projective Invariants for Constant Time Library Indexing in Model Based Vision , 1991 .

[2]  Jonathan J. Hull.  Document Image Matching and Retrieval With Multiple Distortion-Invariant Descriptors , 1995 .

[3]  Benoit Huet,et al.  Cartographic indexing into a database of remotely sensed images , 1996, Proceedings Third IEEE Workshop on Applications of Computer Vision. WACV'96.

[4]  Haim J. Wolfson,et al.  Geometric hashing: an overview , 1997 .

[5]  David S. Doermann,et al.  The detection of duplicates in document image databases , 1997, Proceedings of the Fourth International Conference on Document Analysis and Recognition.

[6]  David S. Doermann,et al.  The Indexing and Retrieval of Document Images: A Survey , 1998, Comput. Vis. Image Underst..

[7]  Jan Flusser,et al.  Point-based projective invariants , 2000, Pattern Recognit..

[8]  Majid Mirmehdi,et al.  Recognising text in real scenes , 2002, International Journal on Document Analysis and Recognition.

[9]  David S. Doermann,et al.  Progress in camera-based document image analysis , 2003, Seventh International Conference on Document Analysis and Recognition, 2003. Proceedings..