Title: Retrieving similar images: using TriCLR+CC-CDLBP features extraction algorithm

Authors: P. John Bosco; S. Janakiraman

Addresses: Department of Computer Science and Engineering, Pondicherry University, Kalapet, Pondicherry, India ' Department of Banking Technology, Pondicherry University, Kalapet, Pondicherry, India

Abstract: The colour-based operator extracts the visual information of an image by analysing the values of neighbouring pixels. Both single and multiple colour features can be derived from various colour attributes, which can serve as image feature descriptors in certain applications. This paper presents a novel feature representation method, TriCLR + CC-CDLBP, which integrates the primary tricolour (TriCLR) approach in colour spaces such as RGB, YCbCr, and L*a*b* with the Change the Central Pixel and Change the Direction (CC-CDLBP) method in LBP. This approach leverages multiple colour features in different colour spaces to enhance image feature representation. A hybrid algorithm is designed to improve the performance of similar image retrieval by applying the primary Tricolour (TriCLR) method to individual channels R, Y, and L* incorporating the CC-CDLBP method to refine feature extraction. By combining TriCLR + CC-CDLBP, our method introduces a robust hybrid approach that outperforms existing image retrieval techniques.

Keywords: colour features; RGB; texture features; CBIR; multiple features; local binary pattern.

DOI: 10.1504/IJCAT.2025.148171

International Journal of Computer Applications in Technology, 2025 Vol.76 No.1/2, pp.115 - 132

Accepted: 20 Mar 2025
Published online: 27 Aug 2025 *