Retrieving digital images from a JPEG compressed image database

Abstract

In this paper, we propose a new method of feature extraction in order to improve the efficiency of retrieving Joint Photographic Experts Group (JPEG) compressed images. Our feature extraction can be done directly to JPEG compressed images. We extract two features, DC feature and AC feature, from a JPEG compressed image. Then we measure the distances between the query image and the images in a database in terms of these two features. Our image retrieval system will give each retrieved image a rank to define its similarity to the query image. Furthermore, instead of fully decompressing JPEG images, our system only needs to do partial entropy decoding. Therefore, our proposed scheme can accelerate the work of retrieving images. According to our experimental results, our system is not only highly efficient but is also capable of performing satisfactorily.