Id Card Number Identification Based on Artificial Neural Network

摘要

His article to include Arabic Numbers, photographs, and a number of special symbol scanning character image as the object of identification, on the key technology of the id card number character recognition research, analyzes various factors that influence efficiency of identification, the effectiveness of the various modules in the process of character recognition algorithm are key research and elaborate in detail, finally complete the character recognition system based on improving the algorithm design, implementation and testing. The main tasks include the following: First, detailed several image the binarization processing methods, compare the treatment effect. Uneven distribution of the target area and the background area, difficult to isolate the problem, proposed and implemented improvements based on local iterative Otsu algorithm to verify its validity by experiment. Second, brief introduction to the feature extraction methods for character recognition. Back-propagation neural network research on the application of a wide range of error, the overall adaptive momentum term and the learning rate for BP network slow network convergence, easy to fall into local minimum defects such improvements, improved classifier and experimental verification of its superiority. Third, complete the entire character recognition system design, implementation, and testing. The actual test results show that: The proposed algorithm, and design of the system is feasible and effective, while the entire system algorithm is able to meet the requirements of quick identification.