Specific Image Content, Automatic Identification And Filtering Technology

Content-based image auto-recognition and auto-filtration is the hotspot of image process, the study of which is very important in theory and applications, especially in sensitive information protection. The system and physics models are designed to recogmze and filter ROI automatically.According to the actual requirements, four modules, image pre-processing, vision features extraction, ROI location and filtration, are analyzed. The algorithms are researched to extract and match the vision features, such as color, shape and space. Mathematical Morphology is used in design of the ROI accurate location and filtration algorithms.Simulation is made to validate the feasibility of the system design and optimize the algorithms. The experiments and related results show that the system designed by this paper has the excellent effect.