Multi-Pose-and-Expression Face Image Retrieval Based on Partitioned Fractal Coding
QU Xin-yue · Qingdao Haiyang Daxue xuebao · 2004
According to the feature that the domain block is in one-to-one correspondence with the least mean square error (LMSE) in fractal coding, a multi-pose-and-expression face retrieval method is proposed in this paper. The retrieval image is divided into range blocks whose sizes are the same as those stored in the database, then the range blocks are coded in terms of preset domain blocks, which results in an LMSE. Compute the Euclid distance between the above LMSE and the one stored in the database. Average the Euclid distances of all the range blocks in the retrieval image, then the images whose average Euclid distances are smaller are selected as the retrieval images. The experiments prove that the images stored in the database can be retrieved rapidly and exactly.