An image retrieval method based on texture features of object region

An image retrieval method based on texture features of object region is presented in this paper, according to Otsu algorithm and local binary pattern (LBP) algorithm. First, it converts color image from RGB color space to gray space. Then the gray image are segmented into object region and background region by using Otsu algorithm and the texture features of object region are extracted by using LBP texture descriptor. Finally, the procedure of feature match is carried out and the retrieval results are listed to user by the descending order of similarity. Experimental results show that the proposed algorithm can effectively describe the texture feature of an image and achieve higher rates of precision and recall compared with the commonly used texture feature algorithms such as gray level co-occurrence matrix and traditional LBP texture descriptor.