Shape descriptor for binary image retrieval

Content-based Image Retrieval (CBIR) system search the images according to the image content such as color, texture, shape and others prominent features from the image. Among these features, shape can give visual structure and appearance of the image. In this paper, we propose a system for binary image retrieval, using Zernike moment, angular radial histogram together with horizontal and vertical histogram of the image. We focus on region-based shape approach and features are extracted from the image. And then, similarity matching process between query image and images in the database is carried out. As a result, the extracted images are shown to the user according with the rank value. Some experimental results by using MPEG-7 CE Shape-1 Part-B dataset are presented.