Based On Artificial Neural Network Handwritten Digital Pattern Recognition And Classification

With the rapid development of science and technology, the processing speed of computeris more and more quickly, using a computer to solve real-life problems is a hot spot ofresearch direction.The handwritten number recognition technology have important theoreticalsignificance and broad prospect of application both in scientific research or in daily life, is themore important research direction of pattern recognition and classification. The handwrittennumber recognition technology involved in many classic problems of the pattern recognitionand classification,such as the feature extraction, classifier design and the choice of samples,and so on, how to improve the accuracy and efficiency is also the scientific research goal ofthe scholars.With the capabilities of parallel computing,fault tolerance,self-learning and classificationand identification, Artificial Neural Network is paid full attention and used in the field ofpattern recognition.In the article, Classifier design adopts the artificial neural networkalgorithm, identification system is established. The learning samples are firstly preprocessedincluding Noise removing, binarization, thinning,and the normalized and so on. Secondly itspixel features are extracted,and the principal component analysis method was adopted toreduce the number of dimensions for saving the time of the training sample.Finally throughthe classifier deal with numerical attributes of the extracted characteristics of sample andclassify to recognise what it is.In the article, Neural network classification and recognitionsystem is established by means of the artificial neural network algorithm,.such as the BPneural network, RBF neural network, probabilistic neural network and so on and eachclassifier’s recognition effect was maked a detailed comparison and analysis.Finally,the paper concluded were related,and proposes future research priorities anddirections.