The Study Of Image Classification Recognition And Object Tracking Based On Feature Extraction And Neural Network

The image classification recognition and object tracking have broad application prospect in the field such as the military, industrial process control, medical research and traffic monitoring, etc. Thus ,the introducing of Neural Network injected new energy to such study.The basic principle of image recognition and object tracking are concisely introduced, and then the BP algorithm which is commonly used is emphasized on.In the same time, the preprocession of image data, the selection of training samples and the design of matching function have been discussed ,the methord of BP network design is introduced as well.The wavelet moment is introduced and cowork with the neural network in the field of image classification recognition. The result of the simulation experiments indicates that this method is feasible and robust. Finally, the information entropy is intrduced and cowork with the neural network in the field of object tracking. It is used in object tracking under complicated background, corresponding tracking algorithm is designed, and the result of the simulation experiments shows this method has good performance.