Region based image retrieval using ranking concept of local binary pattern

Abstract

This paper proposes an efficient content based image retrieval using features derived from regions. The regions are sub divided in to sub regions. On each sub region local binary patterns (LBP) are derived. Each sub region is assigned a unique code if it forms a fundamental unit i.e. a uniform LBP (ULBP). The present paper derives a unique number for each region called as rank of region. The rank of the region is computed based on the number of ULBPs each sub region derives of that region. The present paper derived rank based region ULBP cooccurrence matrix (RR-ULBP-CM). The gray level cooccurrence matrix (GLCM) features are evaluated on RR-ULBP-CM for efficient image retrieval and experiments are performed using Corel data base. The proposed RR-ULBP-CM method is compared with the existing methods and results establishes the efficacy of the proposed method.