Wavelet-Based Multiresolution Histogram for Fast Image Retrieval

One of the important task in most of Content Based Image Retrieval(CB1R) system is similarity matching. Similarity matching requires distance computation of feature vectors for each candidate image in image database. Conventional algorithms based on exhaustive search are highly time consuming and inefficient. With rapid increase in database size, there is a growing need of a fast and efficient retrieval system. Multiresolution data-structure based approach provides a good solution to above problem hut there is still scope for improvement. In this paper we propose a wavelet based multiresolution data-structure algorithm for faster image search. The proposed approach reduces the number of computation by around 50% over multiresolution datastructure algorithm. In the proposed approach we reuse the information obtained at lower resolution levels for similarity matching at higher resolution levels. This algorithm also saves disk storage space by ahout 50% over multueolution data-structure approach. The proposed approach can he easily combined with existing algorithms for further performance enhancement. In this paper we use the proposed approach to match similarity between luminance histograms for image retrieval. keywordsContent based image retrieval, fuzzy histogram, multiresolution histogram, wavelet histogram, feature matching.