Region-Based Image Retrieval with Scale and Orientation Invariant Features

In this paper, we address the problem of image retrieval when the query is in the form of scaled and rotated regions of images in the database. The solution lies in identifying points that are invariant to scaling and rotation and determining a robust distance measure that returns images that contain the query regions. We use the Harris-Laplacian detector to detect the interest points which are then matched with similar points in the image database using a novel fuzzy distance measure. Images with closely matching interest points are further refined using a cross-correlation measure that results in the final set of retrieval images. Experimental results show the effectiveness of the proposal image retrieval strategy.

[1]  C. Schmid,et al.  Indexing based on scale invariant interest points , 2001, Proceedings Eighth IEEE International Conference on Computer Vision. ICCV 2001.

[2]  Adam Baumberg,et al.  Reliable feature matching across widely separated views , 2000, Proceedings IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2000 (Cat. No.PR00662).

[3]  Gerald Sommer,et al.  A Fuzzy Logic Algorithm for Dense Image Point Matching , 2001 .

[4]  Cordelia Schmid,et al.  Image Retrieval , 2001 .

[5]  David G. Lowe,et al.  Object recognition from local scale-invariant features , 1999, Proceedings of the Seventh IEEE International Conference on Computer Vision.