Uncertainty Based Sampling Approach for Relevence Feedback in Content Based Image
M. G. M's, Archana M. Rajurkar · 2014
Interest in the digital images has increased enormously over the last few years, but the process of locating a desired image in such a large and varied image collection becomes very difficult. Traditionally text in different languages is used for efficient retrieval of images; it has several drawbacks such as language constraint and subjectivity of human perception. Content-based image retrieval is a technique which uses visual contents such as color, texture and shape to search images from large image databases according to user's interest. Color is the most commonly used feature for content based image retrieval. The major drawback of usual color histogram based method (binning method) is that, it does not take image color distribution into consideration and inflexibly partition the underlying color spaces into a fixed number of bins. In this paper we propose a moment-preserving technique based on binary quaternion space for feature extraction. It aims to extract color features according to the image color distribution that significantly reduces the distortion incurred in the feature extraction process. It is observed that minimizing the distortion incurred in the feature extraction process of proposed color distribution based approach can improve the accuracy of retrieval. Our experimental results show that the proposed extraction methods can enhance the average retrieval precision rate by a factor of 25% over that of a traditional color histogram based feature extraction method. It is also observed that, this technique effectively reduces the average retrieval time. Further, in this paper we propose a novel probabilistic approach for relevance feedback based on the theory of uncertainty based sampling. It significantly enhances the retrieval precision and simplified the task of query refinement and improves usability of a system.