Extraction Of Salient Region In Image And Application In Content Based Image Retrieval

With the development of information technology, images become a major carrier of information. But more and more image data cause inconvenience in manual processing and browsing. In image analysis and understanding, the salient region is usually only a small part of an image. Here the salient region can be regarded as the main object in an image, which is the interesting region that human vision can quickly focus on. In the thesis, based on a comprehensive analysis on the gray-scale information, color information, and orientation information of an image, we propose an approach to simulate the principles of human vision to extract the salient region in an image.The image retrieval technique refers to finding interesting or desired images in VLIDBMS (Very Large Image Database Management System). Currently, many image retrieval tools are based on textual information related to images and require a large amount of human labeling. CBIR (Content-based image retrieval) analyzes the image content itself, and finds...