Content-based Image Retrieval Research

Along with the development of computer vision, Multimedia and database technology, image information, a main kind of multimedia information, is widely used in many fields. In order to retrieval the images which user want to find quickly and accurately from large amount of images, so Content-based image retrieval (CBIR) technology has been studied a lot currently.Based on the understanding of the content on CBIR and its related fields technology, this paper sets focus on the feature extraction of image, and the main research content of this paper include the following aspects.Firstly, since HSV color histogram can't describe spatial information effectively, so this paper has made some modification on the retrieving algorithm of color feature. In the color indexing approach using the dominant colors of multi-resolution partitions uses an overlapping rectangle multi-resolution partitions way, and makes up the deficiency of traditional multi-resolution partitions ways; in the approach of Color Coherence Vectors, this paper chooses the HSV color model. By the experiment, this paper validates the feasibility of the above extracting color feature algorithms.Secondly, in order to design an effective mechanism of relevance feedback using in image retrieval system based on color feature, this paper has studied and realized a color feature re-weighting approach for relevance feedback. With plentiful experiments, it is proved that the efficiency can be enhanced by using this method.Thirdly, in order to use Multiple Features in image retrieval system, this paper has studied the image retrieval method using color feature and texture feature. To represent the color content of an image, the color indexing approach using the dominant colors of multi-resolution partitions is used. To represent texture feature, gray co-occurrence matrix is computed. Through the image retrieval experiment, better image retrieval performance can be achieved by combing two kinds of features.At last, this paper has realized an image retrieval system, and this system can help users retrieval images they need more effectively from a large image library.