Research On Image Retrieval Based On Multi Feature Fusion And SVM

With the rapid development of Internet technology,multimedia technology and the arrival of the big data era,the image data on the Internet is in the trend of explosive growth.People's daily life has been filled with a variety of image information.How to get the image information quickly and accurately from a large amount of image data becomes an urgent demand.Content-based image retrieval technology is the focus and direction to solve this problem.However,the performance of content-based image retrieval is affected by the semantic gap between the low-level features and upper-level understanding of images.In this paper,two aspects of image retrieval technology are researched:the first·part of the image retrieval based on multi feature fusion;the second part of the image retrieval based on SVM relevance feedback.The main work is as follows:(1)An image retrieval algorithm based on multi feature fusion is proposed.Firstly,the method of image feature extraction is studied,the global image features GIST is extracted from the overall description of an image.Then the similarity measure method is used to find the matching image in the database.After that,the SIFT features of scene similar matching images are extracted as the local features of the image.As the SIFT features of the image are some feature points,so the point matching algorithm based on BBF search algorithm is used in matching.Finally,the search results are returned according to the matching numbers of SIFT feature points between the query image and the scene similar image.The image retrieval is realized by the combination of the GIST feature and the SIFT feature based on the idea of the global recognition first.The retrieval performance of multi feature fusion algorithm and single feature algorithm is compared by experiments.The experimental results show that the performance of the multi feature fusion algorithm is better than the single feature retrieval algorithm.(2)On the basis of traditional content-based image retrieval,image retrieval based on SVM relevance feedback is studied.The idea of machine learning is added to image retrieval.After discussing the relevance feedback technology and support vector machine technology,the traditional method is used for image retrieval.Then,the retrieval results consistent with user search intent are obtained according to the relevance feedback technology based on SVM.The experimental results show that the image retrieval based on SVM relevance feedback is effective and has certain practical value in image retrieval.