A Content-based Image Retrieval Scheme in Jpeg Compressed Domain

Nowadays, a large number of images are compressed in JPEG (Joint Photographic Experts Group) format. Therefore, content-based image retrieval (CBIR) for the JPEG images has attracted many people’s attention and a series of algorithms directly based on the discrete cosine transform (DCT) domain have been proposed. However, the existing methods are far from the practical application. Thus, in this paper, a new image retrieval scheme for JPEG formatted images is presented. The color, spatial and frequency (texture) features based on the DCT domain are extracted for the later image retrieval. It doesn’t require decompressing the images but directly retrieving in the DCT domain. Thus, compared with the spatial domain based retrieval methods for JPEG images, the computation complexity can be greatly reduced. In addition, this retrieval system is suitable for all color images with different sizes. Experimental results demonstrate the advantages of the proposed retrieval scheme.