A new segmentation architecture for texture matching using the LBP method

In this paper, we propose a texture matching approach based on a new decomposing architecture. First, the main window starting from the converging point a to the bottom-right corner of the image is considered. For each iteration, the main window´s size is reduced by a distance d and other windows with the same main window´s size are generated if possible. The Local Binary Pattern (LBP) technique, which is invariant to monotonic grey level changes, is used for statistical texture feature extraction for each window. Synthetic images are used in the experimentations using the proposed system and some results are shown in the last section.