A Novel Approach Based on Logistic Regression and Bayesian for Relevance Feedback in Content-Based Image Retrieval
Jun Kong, Xuefeng Wang, Zhen Liu, Xiaohua Zhang, Jingxia Cui, Jingbo Zhang · 2008
This paper presents a new relevance feedback (RF) method for image retrieval in content-based image retrieval (CBIR). The main conception of the method gives two aspects: First logistic regression adjusts the weight of each element in features extracted from the images in database with the preferences of the user. Then following a Bayesian methodology, which yields the posteriori of the images in the database and used to show to the user a new set of images. The retrieval system is repeating until he/she is satisfied or the target image has been found. Experimental results show the superiority of the proposed method.