Research And Implementation Of Handwritten Numeral Recognition Technology

With the advent of the information age, people for information processing requirements more stringent and rapidly. Handwritten digit recognition technology, the information for future development plays an important role in the field of optical character recognition.At the same time,as we all know that handwritten character recognition is a very important branch,which has been widely applied in real life. For example, in identifying bank notes, in accordance with handwritten identification numbers; in the mail sorting process, according Zip identify automatic sorting, save trouble and effort. In addition, handwritten character recognition pattern recognition for the development of other Chinese character recognition, face recognition and other related areas, but also has a great role in promoting. Especially with the advancement of technology, the application is now smart phones, tablet computers and other electronic equipment is very popular, more and more people are accustomed to using handwritten Chinese characters or numbers. Therefore, the study of handwritten digital identification technology, not only has important theoretical academic significance, as well as a wide range of realistic value.Neural network is a hot topic in recent years, too, because it itself has a high degree of parallelism, non-linear, high redundancy features to solve complex problems is increasingly becoming the preferred conditions. Neural networks for knowledge representation and storage of information processing becomes easy to make, in addition, has its own self-learning, self-organizing capacity for complex models with a greater advantage. This paper describes the background and significance of the character recognition technology to analyze the research status, then studied handwritten digit recognition in some of the key technologies in-depth study of its principles and methods, the proposed recognition technology to improve the accuracy of identification, to the requirements of practical applications. Meanwhile, the artificial neural network technology for the understanding and use, BP neural network is one of the largest hotspot network of artificial neural network, BP neural network used in the system for processing, using HOG feature extraction methods, the training data and test data were many experiments, made some experimental results.In this paper, the main work of this paper in the following areas:1. In the study of a variety of basic handwritten numeral recognition feature extraction, using HOG feature extraction method, you can quickly and accurately handwritten digital image feature extraction.2.Samples of BP neural network training and testing, through a lot of experiments on handwritten numeral recognition analysis and study.3.The experimental results were analyzed systems and propose an improved method, the next step would be the relevant part of the work carried out in-depth research.