Visual Cognition Based Image Searching And Recognition Key Tecnology

The challenges to image pattern recognition lie in such the various requirements for image retrieval and understand in high rank task of image process as well as the complex, hard described image content, and the gap between the digital array of image expression and the conceptual information universally accepted by human being. On the basis of image information from visual perception, a model integrated with bottom up visual searching and top down semantic determining is proposed in this dissertation, establishing a perception contact between image areas or image blocks and image semantic or image categories. In addition, image neuron index and cognitive information rough set methods arc emploied. which set up the indices for the interesting contents, and carry out the recognition law discovery as well as the collaborative search and identification under the driving of task events.Through the scanning and searching for image areas as well as visual contact learning of image blocks near the focus of attention, the perception contact of visual feature and semantic feature of image blocks is built up. making visual perception behavior of image and cognitive behavior in image understanding into a whole. The model blends two aspects of visual perception and semantic explanation in image understanding, which conforms to the human cognitive law.Combination of bottom up feature map with top down semantic predications establishes the basic framework contact between images and their semantic classes. Based on semantic association, a semantic representation for a scenic category is formed so that users can organize the similar images in perception together to form perceptional context and from which users can explain and mark images without need to give the connotation description for images.In users’ feedbacks to query results, positive correlation images are reserved while the negative correlation images are removed. Under the basic frameworks, through multi-level semantic group link and multi-framework coordination, the images of original framework are well reused, and the correlation between the visual cognition, semantic understanding and query cases is effectively realized, which meets the requirement of flexible retrieval of images.