With neural network research and the increasing depth means of achieving maturity, Neural networks are widely used in pattern recognition problems, in particular for the various image recognition. Among them, BP neural network is the most widespread application of a multi-supervised learning neural network It is widely used pattern recognition, system identification, prediction, and control applications. BP sample input and output as a highly nonlinear mapping, the use of the gradient descent algorithm, Operators using iterative solution to the corresponding value in learning and memory problems. According to the design of the same image characteristics moment, BP neural network theory, MATLAB environment in the completion of the BP neural network training process; BP neural network to prepare the dynamic link library, and the VC environment called for the link to realize some of the basic images (rectangular, round, etc.) of pattern recognition. The image processing software packages (including image enhancement, image segmentation, image feature extraction, Image pattern recognition, etc.) of a component.完成!!!
|
