Handwritten English Character Recognition System

Characters are important carriers for human information communication. With the development of the computer and information technologies, the processing and recognition of character information in computer has become very important. Now handwritten character recognition is a very important and active research field in pattern recognition. It involves pattern recognition, digital image processing, digital image processing, artificial intelligence and so on. It is a comprehensive technology and has a very broad application background and giant market demand. Thus it is of both theoretical and practical significance.There are many problems which involve handwritten character deformation randomness in off-line handwriting character recognition. At present, the off-line handwritten character recognition technology is immature and still need research. In this paper, we try to combine the neural network with pattern recognition technology. With the help of advantages, which involve good tolerance for error, strong sorting ability, strong parallel handing ability and strong self-learning ability of neural network, we develop a method for off-line handwritten character recognition. Image pre-processing and normalization are carried out in the paper. After the operation, we start to extract image feature of handwritten characters and take the feature vector which has just been extracted as the input vector of neural network recognition input vector. We add momentum and change learning rate in BP neural network and with the help of simulated annealing, we achieve the optimization and improvement of BP ( neural network. Then training and recognition of neural network are fulfilled. Our algorithm can process vast data quickly and precisely, and the recognition effect is fairly good.The paper finally build a neural network handwritten character recognition system in Visual C ++ platform. The system has optimized the structure, robust and generalization ability. It achieves high accuracy of recognition .