Color Texture Description with Diagonal Local Binary Patterns Using a New Distance Metric for Impressive Image Retrieval

Abstract

With the reliability and dependence of the world over the digitalization of information in all spectrums of professional and personal life, the volume of digital data is increasing exponentially. With such a huge database, retrieving information, and images from the storage media pose serious challenges. In this paper, a diagonal local binary pattern for color images (DLBPC) and a new distance metric are proposed for the retrieval of data with high accuracy and speed. The descriptor employs the deviceindependent L*a*b* color space. The discriminative quality of DLBPC is enhanced by combining local features with the (color moment) global feature of the image for effective image retrieval. To assess the efficacy, the suggested system was implemented on challenging datasets: Wang -1K and GHIM-10K. The analysis findings demonstrate the efficacy of the suggested system over other reported systems.