International Journal of Computer Vision
Combining supervised learning with color correlograms for content-based image retrieval
MULTIMEDIA '97 Proceedings of the fifth ACM international conference on Multimedia
Hi-index | 0.00 |
Each content-based image retrieval (CBIR) method using color features includes its limitations to be applied to its own application areas. We analyze that the limitations are mostly due to the adoption of first-order or second-order relations among color features from a given image as its index. In this paper, we propose a new CBIR method based on a third-order color feature relations. This new method shows robustness in retrieving geometrically transformed images and can be applied in various areas.