Abstract: This paper proposes a survey on Content Based Image Retrieval via various features. Related work is the foundation for understanding and improving more knowledge regarding precise field. Content Based Image Retrieval is a technique for extracting features of images like color, shape, texture, blob detection, edge detection, contour detection and etc. Once features are extracted from a query image then same procedure will be used for extracting features of database images. For comparing features of both images Euclidian distance can be used to produce the best result. The key part of retrieval system is feature extraction. Until now, the only way of penetrating these collections was based on keyword indexing, text, or just by browsing. In this paper we survey some practical aspect of current CBIR and color histogram, texture, and shape for exact and efficient CBIR after doing the study of related works.

Keywords: CBIR, ABIR, Precision, Recall.