A noble color-texture hybrid method for content-based image retrieval

In this paper, a color and texture hybrid framework CCLBM is proposed for content based image retrieval. The purpose of this study was to construct a method for image retrieval that is relatively simple in computational complexity and still maintains good retrieval results. The proposed color-texture hybrid method CCLBM combines Color Coherence Vector for color information and Local Binary Pattern for texture information. The method was tested and benchmarked on COREL-1K and COREL-5K image databases in terms of Average Retrieval Precision (ARP) and Average Retrieval Rate (ARR). The results show that CCLBM scores 82.52% and 67.13% in terms of Average Retrieval Precision (ARP) in COREL-1K and COREL-5K databases respectively (for top 10 similar image retrieval).