A Review: Color Feature Extraction Methods for Content Based Image Retrieval

Abstract

For more than a decade Content Based Image Retrieval is topic of interest for researcher. The three primitive visual features, namely, color, texture and shape refers to the term ‘content’ in content based image retrieval. Although visual features cannot be completely determined by semantic features, but still semantic features are used because they are easier to integrate into mathematical formulations. Therefore good visual feature extraction is one of the important task for representing image compactly. Among the visual features, colors is the most vital, reliable and widely used feature. This paper reviews various methods, namely Global Color Histogram, Histogram Intersection, Image Bitmap, Local Color Histogram, Color Correlogram, etc., employed to extract the color feature. This paper briefly elaborates these different methods of color feature extraction and then presents a comparative study for selection of these methods in various applications.