An Object Recognition And Segmentation Method Based On Boosting And PDE

In recent years, object recognition and segmentation methods have become one of hot spots in computer graph model processing, and have attracted more and more attentions. In order to get better results, we should extract features for the recognition object, and find recognition and segmentation method for the object.In this paper, several mainstream object recognition and segmentation methods are summarized carefully in theory, we find a new recognition and segmentation method. The contributions of our work include the following contents:(1) In order to do further research, we have set up the image model database.The database include the two following parts:car image database and screen image database.(2) We have analyzed many mainstream feature extraction methods seriously in theory, and evaluated the advantages and disadvantages of various methods, through experimental results to find a feature extraction method based on fragments.(3) Several recognition methods are analyzed carefully, we find a recognition method based on boosting. The algorithm is a new recognition method. Compared with the traditional algorithms, the experimental results show that this algorithm has high precision, and strong robustness.(4) Many image segmentation methods are summarized seriously in theory and image segmentation methods based on PDE are studied. Compared to tradition image segmentation approaches, the sgmentation methods based on PDE have fast segmentation speed, and high accuracy.(5) Many object recognition and segmentation methods are summarized carefully, an object recognition and segmentation method based on Boosting and PDE is proposed. Finally, we establish a new recognition system for images. Lots of validation experiments have been carried out for the system. We have analyzed the experimental results seriously, the experimental results show that our system is very effective and has high recognition precision.