Research Of Image Classification And Retrieval Techniques Based On SVM

With the popularity of the Internet and multimedia technology, the techniques of image classification and retrieval technology has recently become a hot needs and research. This paper studies the content-based image retrieval (CBIR) technologies, including color and texture of the visual characteristics of extraction and description algorithm, the similarity measure, normalized, and on the basis of further studing on the histogram, we put forward the improved gray area histogram algorithm. Then through the research of the principles of support vector machine (SVM) to achieve a SVM-based image classification and retrieval of experimental system, and we have made some modifications to the related feedback (RF) algorithm. Experiments show that on the basis of pre-classification SVM-based image databases,using the integrated feature extraction algorithms with RF approach produced by this paper gets more improvement on the effect,and can better meet the needs of users.