An Image Retrieval Method Based on Relevance Feedback and Collaborative Filtering

With image retrieval becoming increasingly important, the application of relevance feedback and content-based image retrieval technique has become a research hotspot. For the shortcomings in many of the existing image retrieval methods where there is the information of relevance feedback which is not be fully saved and used, and has poor accuracy and flexibility, an image retrieval method based on relevance feedback and collaborative filtering was proposed. It was discussed that how to improve efficiency of relevance feedback, reduce the number of interactions through analysis of historical data of relevance feedback and speed up the feedback process. Finally this method was compared with the existing retrieval methods. The experimental results showed that this method had significant improvement in retrieval effectiveness, which can effectively improve the rate of identifying all and precision rate.