Significance of the Background Image Texture in CBIR System

Abstract

The Content Based Image Retrieval (CBIR) using local texture feature is extensively studied for decades. The local texture features such as LBP, LTP, LDP, and other derived local texture features uniquely describes the content of an image. The performances of the local texture features with different classifiers are well exposed by the researchers. The purpose of this paper is to explore the effect of the LBP texture feature due to simple and complex background in the image. The experiments are done with the Wang database using various distance measures. The results are promising and it shows the LBP feature is biased by the complex background and it affects the performance of the CBIR system.