Object Recognition for Content-Based Image Retrieval

Early CBIR systems searched for images based on their global appearances. This kind of approach requires users to provide a sample image first, and the image retrieval task is performed based on low-level image features, but not on the actual objects found in the image. So,user-friendly CBIR system must operate on the same semantic level as its human users. Since humans classify images according to their objects and concepts, the system must have the ability to recognize object and concept classes in order to automate the process of image annotation. However, object recognition is still an open field for computer vision research, and most successful object recognition systems only work for some particular objects or object classes.