Storage of Images and Video in Database Systems

We consider the problem of constructing indices to image and video data analogous to those used in classical database systems, in order to speed up the retrieval of such data from physical storage (either local or remote). To achieve this goal, we propose new coding techniques which combine a wavelet representation, embedded coding of the wavelet coe cients, and segmentation of image-domain regions in the wavelet domain, to generate a bitstream in which each image region is encoded independently of other objects, and without explicitly storing shape boundary information. Our simulation results show that the new proposed algorithms achieve coding performance which compares favorably, both perceptually and objectively, to that achieved using state-of-the-art image/video coding techniques which do not support indexing. Our results establish the feasibility of physical database designs for image and video data in which the compression performance of state-of-the-art coders needs not be compromised, thus allowing low data access times, a major factor a ecting the performance of Database Management Systems (DBMSs).