Content-Based Image Retrieval Using Grey Relational Analysis

In general, there are two fundamental issues in content-based image retrieval (CBIR): content representation and similarity matching. Once the set of feature vector representing the content of images is determined, the question of how to combine the individual similarities according to each term of the vector to form a global similarity must be addressed. In this paper, the low frequency DCT coefficients that transformed from YUV color space are used to form the feature vector of an image. Then, a new similarity measure based on the grey relational analysis is introduced to evaluate how close they are between two DCT sequences. Experimental results show that it is an easy and practical method in the field of CBIR.