Realization of Handwritten Numeral Recognition System Based on PNN with MATLAB
Sanping Li · Journal of Military Communications Technology · 2005
MATLAB software has been used extensively in engineering technology fields due to its strong function. The classifier of Bayes based on least error probability has the best classifying effect in performance. The main trait of neural network lies in the better performance of error contain, the ability to classify, the power of parallel disposal, and the auto-learning ability. The PNN has the better synthesis classifying effect in practical application for integrating the predominance of the Bayes classifier and neural network. This article realized a software of real time handwritten numeral recognition system in the environment of MATLAB 6.5. The classifier was PNN in this system. Experiment testified the good performance of the software with better extensibility and currency. It also offered a simple effective emluator for the study of handwritten numeral recognition.