Research Of Image Retrieval Algorithms Based On Color

With the explosion of multimedia data, the traditional image retrieval method by text couldn't meet people's demand of more accurate retrieval results any longer. Therefore, the content-based image retrieval (CBIR) has been researched to achieve more accurate results. CBIR uses image visual features to represent image and perform retrieval. Color feature is applied most widely in image retrieval systems. In this paper, we choose several common CBIR algorithms based on color to analyze their robustness to the characteristics of images. We test 9 kinds of images for the algorithms. From the experiment performance, we evaluate the adaptability of the algorithms under different kinds of images.