Block-based, memory-efficient JPEG2000 images indexing in compressed-domain

The contribution of this paper is the development of a fast, block-based JPEG2000 image indexing system in the compressed domain which achieves high memory efficiency. By high memory efficiency we mean the entire wavelet-decomposed coefficient matrix never needs to be buffered. The feature extracted is the variance of each rectangular block in the compressed domain. The extraction of this feature is of very low complexity. Retrieval performance on the VisTex image database has shown its effectiveness.