• Dr. M. NESTER JEYAKUMAR

ABSTRACT: Content Based Image Retrieval (CBIR) is associated with the recovery of identical images from image repositories, employing the feature vectors obtained from images. These feature vectors provide the global definition for the visual content that exists in an image, for instance, texture, colour, shape, and spatial associations between the vectors. In our recent work the Complete Local Spatial Distribution Pattern (CLSDP) was proposed to successfully represent the spatial distribution image feature. In this paper, defining the feature vectors employing the Local Binary Pattern (LBP) operator is proposed. In order to improve the retrieval performance, we will also take into account a feature coding method based on Circular Shift Right (CSR) to make CLSDP with rotation invariance and scale invariance. Circular Shift Right Linear Network Coding (CSR-LNC) is proposed to CLSDP based on circular shift right with rotation invariance and scale invariance. CSR-LNC was performed in order to determine the optimum CLSDP and here circular shift right is proposed for the general definition of image feature vectors with rotation invariance and scale invariance. CLSDP with CSR-LNC is proposed for CBIR applications which capture the directions in a local region. To compare its performance, this proposed CLSDP with CSR-LNC technique is applied to a Corel 1K Database and is compared with the CLSDP, and LBP. Result shows that the proposed technique achieves good retrieval rate. The outcomes reveal that CLSDP with CSR-LNC technique offers outstanding performance in comparison with the available algorithms in terms of both mean precision and recall.

INDEX TERMS: Feature extraction or construction, Content Based Image Retrieval (CBIR), Feature representation, Complete Local Spatial Distribution Pattern (CLSDP), Circular Shift Right Linear Network Coding (CSR-LNC).

"CIRCULAR SHIFT RIGHT LINEAR NETWORK CODING (CSR-LNC) WITH LOCAL PATTERNS FOR CONTENT-BASED IMAGE RETRIEVAL", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.5, Issue 12, page no.83-88, December-2018, Available :http://www.jetir.org/papers/JETIR1812812.pdf

"CIRCULAR SHIFT RIGHT LINEAR NETWORK CODING (CSR-LNC) WITH LOCAL PATTERNS FOR CONTENT-BASED IMAGE RETRIEVAL", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.5, Issue 12, page no. pp83-88, December-2018, Available at : http://www.jetir.org/papers/JETIR1812812.pdf