Effective image retrieval using dominant color descriptor and fuzzy support vector machine

Image retrieval is an active research area in image processing, pattern recognition, and computer vision. Relevance feedback has been widely accepted in the field of content-based image retrieval (CBIR) as a method to boost the retrieval performance. Recently, many researchers have employed support vector machines (SVMs) for relevance feedback. This paper presents a fuzzy support vector machine (FSVM) that is more robust to the four major problems encountered by the conventional SVMs: small size of samples, biased hyperplane, over-fitting, and real-time. To improve the performance, a dominant color descriptor (DCD) is also proposed. Experimental results based on a set of Corel images demonstrate that the proposed system performs much better than the previous methods. It achieves high accuracy and reduces the processing time greatly.