Texture Image Database Retrieval Using JPEG-2000 Partial Entropy Decoding

摘要: In this paper, we propose a novel JPEG-2000 compressed image retrieval system using feature vector extracted through partial entropy decoding. Main idea of the proposed method is to utilize context information that generated during encoding/decoding. framework JPEG-2000, current coefficient determined depending on pattern significance and/or sign its neighbors in three bit-plane coding passes and four modes. The contexts provide model for estimating probability each symbol be coded. And they can efficiently describe texture images which have different because represent local property images. addition, our directly search domain without full decompression. Therefore, scheme accelerate work retrieving We create various distortion similarity databases MIT VisTex simulation. evaluate algorithm comparing with previous ones. Through simulations, demonstrate achieves good performance terms accuracy as well computational complexity.