D.Kishore , S.Srinivas Kumar, Ch.Srinivasa Rao
Jawaharlal Nehru Technological University,Vijayanagaram, India

In this paper, gray level co-occurrence matrix, gray level co-occurrence matrix with singular value decomposition and local binary pattern are presented for content based image retrieval. Based upon thefeature vector parameters of energy, contrast, entropy and distance metrics such as Euclidean distance, Canberra distance, Manhattan distance the retrieval efficiency, precision, and recall of the images are calculated. The retrieval results of the proposed method are tested on Corel-1k database. The results after being investigated shows a significant improvement in terms of average retrieval rate, average retrieval precision and recall of different algorithms such as GLCM, GLCM & SVD, LBP with radius one and LBP with radius two based on different distance metrics.

CBIR, GLCM, SVD and LBP.