The Recognition of Handwritten Digits Based on BP Neural Network and the Implementation on Android

Offline handwriting recognition has become one of the hottest directions in the field of image processing and pattern recognition. It can transform any handwriting to plain text file and has been widely used in cheque recognition, mail sorting, reading aid for the blind and so on. In this paper, we attempt to recognize handwritten digits with feature extraction by Back Propagation(BP) neural network. The MNIST database of handwritten digits is applied to train and test the neural network. In addition, we introduce the Principal Component Analysis (PCA) for feature extraction which can improve the performance of the neural network and immensely shorten the training time. On the other hand, we make comparison of the recognition rate among three methods: neural network method, thirteen features method and Fisher discriminant analysis method. Finally, we succeed in porting these algorithms to Android.