关键词: 基于内容的图像检索, 离散余弦变换, 交流系数分布熵

Abstract:

Introducing the spatial distribution information of alternating current (AC) coefficients into the feature extraction,a new Joint Photographic Experts Group (JPEG) image retrieval is proposed. Firstly，according to the characters of AC coefficients, invariant for the rotation, AC coefficient distribution entropy is presented. Meantime, because of the different effect of each AC coefficient on the retrieval precision, the weight function for the entropy is proposed; thus can avoid the mistaken retrieval and losing retrieval. Not only statistical distribution but also the spatial distribution of discrete cosine transform (DCT) coefficients are taken into account, it is robust for transform, rotation and scale. Experiments show that the new method has a better performance both in retrieval efficiency and effectiveness.

Key words: content-based image retrieval, discrete cosine transform, alternating current coefficient distribution entropy