摘要： 提出一种JPEG图像检索算法，将熵解码后的DC系数差本地二值化，提取DC差向量来体现原始图像中像素的灰度分布。考虑到每个DCT块中的AC系数对图像内容的表征能力，针对每个DCT块选择前9个最大的AC系数，构造AC系数特征向量用于检索。实验结果表明，该算法具有较好的检索效果。

关键词: 图像检索, 离散余弦变换, DC系数, AC系数

Abstract: This paper proposes a JPEG image retrieval algorithm. After entropy decoding, the difference of DC coefficients is becoming binary data. According to the character, the DC vector is introduced to embody the gray distribution of the original image. The AC coefficients of the distribution are taken into account and nine AC coefficients are chosen to extract the AC distribution vector. Experimental results show that the proposed method has a good performance in retrieval efficiency.

Key words: image retrieval, Discrete Cosine Transform(DCT), DC coefficient, AC coefficient

中图分类号: 

• TP391