Evaluation of Feature Extraction Techniques in Content Based Image Retrieval (CBIR) System

The term CBIR refers to the process of retrieving similar images from a large collection of image database. The image retrieval is done on the basis of similarity matching between query image and database images. Different feature extraction techniques are being used to extract features of an image. The important image features are Color, Texture, Shape, Spatial location, edges etc. The CBIR (Content Based Image Retrieval) systems have achieved wide growth in past few years but still there are many limitations in CBIR systems yet to overcome. In this paper, we have evaluated feature extraction techniques based on color, texture and shape information and find that not a single feature vector is good for the retrieval parameters like precision and recall. There is the need of combining multiple feature vectors for increasing the effectiveness and efficiency of CBIR Systems.