Image retrieval based on color, shape and texture

In this paper, we have hybridized the three main techniques of Content Based Image Retrieval (CBIR) in order to increase the efficiency and precision of the images which are retrieved. Firstly, the color quantization of the image is done with the help of ColorCorrelogram Vector. The spatial arrangement of the color pixel is also determined by the colorcorrelogram. Then, the shape and texture of the image is found out by the BDIP(Block Difference of Inverse Probabilities) and BVLC (Block Variation of Local Correlation) which are block-based techniques. The shape of the image is also been predicted by the Prewitt operator. The Euclidean Distance is being used in order to compare the two extracted feature vectors i.e. the target vector and the query image vector. In the end, there are retrieved images which not only have high recall but also have high precision or accuracy. Experimental data shows that this combination of techniques is proving out to be much more efficient and accurate in retrieving the user interested images. The main aim behind the formation of this CBIR is to match with those images which are been selected with human perception.