Effective Distance Measures for Color-Content Based Image Retrieval

Content-based image retrieval systems are the challenging area of research in computer science. A retrieval method which uses color feature is implemented in this paper. Effectiveness of various distance measures such as euclidean distance, histogram intersection, histogram quadratic distance and canberra distance are also implemented to determine similarity between feature vectors. Results are compared to determine performance of each of the distance measure during image retrieval. The experimental results show that histogram quadratic distance measure produced more accurate results for image retrieval as compared to other distance measures.