Handwritten Numeral Recognition Based on BP Neural Network

Handwritten numeral recognition is a hotspot of study for years,and is an especial issue of character recognition.On account of great changes of handwritten font,it is very difficult for the traditional method of recognition to achieve high recognition rate.To counter the complexity and limitation of traditional digital recognition methods,a kind of handwritten numeral recognition method based on BP neural network is proposed.The point feature and stroke density feature for handwritten digits are extracted;then an improved BP neural network is applied to classify handwritten digits by those features.Via experiment,the recognition rate is 94%.Experiments show that the proposed approach has a good effect on handwritten numeral recognition.It not only simplifies the complexity of the traditional recognition,but also increases the accuracy of recognition.