CBIR using Combined Feature Vectors of Column-Wise and Row-Wise DCT Transformed Plane Sectorization

Content Based Image Retrieval is a way of computer viewing technique used to retrieve digital images from a huge database. In this paper we have first calculated the feature vector column-wise and row-wise separately. After this we have concatenated the feature vectors of column-wise and row-wise. To evaluate the performance of the proposed method we have used Precision-Recall crossover point, LIRS, LSRR and LSRI. Sum of Absolute Distance and Euclidean Distance are the two similarity measures used. The column-row wise DCT transformed image is sectorized on the basis of even-odd column components of transformed image with augmentation of zero and highest row components. The proposed algorithm is applied to a database of thousand images. These thousand images are grouped in ten different classes. Performance is evaluated and compared for 4, 8, 12, 16 DCT sectors.