Applying Iterative Logistic Regression and Active Learning to Relevance Feedback in Image Retrieval System

This paper presents a novel relevance feedback algorithm for image retrieval in content-based image retrieval systems based on the Logistic regression model. In order to narrow down the semantic gap between user´s high-level query concepts and the low-level image features, user preferences are added to the algorithm. Based on modeling of user preferences as a probability distribution, the algorithm can calculate the relevance probability of an image belonging to the set of those selected by the user. And it ranks the images according to their probability. The process is repeating until the user is satisfied with the query results or the target image has been found. The problem of scarcity of labeled (training) examples in the feedback process is effectively addressed by meaning of tracking the subset and active learning method. Experimental results are shown that the performance of the retrieval system is greatly improved by the proposed method.