With the development of Internet and multimedia technologies, Images retrieval based on their visual features becomes an important research area in distributed multimedia system. In this paper, we propose a new method for images retrieval by their visual appearance-Multi Scale Gauss-filtering Representation (MSGF). Images are first filtered in several scales with Gauss derivative filters, and the feature vectors are constructed on filtering outputs in several scales. Next, images are matched by their feature vectors, the matching results are scored and returned to user. We develop a fast search algorithm of index structure for local similarity retrieval of images. We then generalize the idea to a wide range of image retrieval by global similarity. Experiment shows that this method is effective in feature extraction and images retrieval by their visual appearance.
