基于QNN的图像特征提取字符识别系统设计

为提高字符识别的正确率与可靠性，将图像处理技术与量子神经网络（QNN）相结合，对基于QNN的字符识别系统进行了研究。采用粗网格特征法对图像的特征量进行提取，同时，为了增强粗网格特征法抗位置变化的能力，在特征提取前，对字符图像进行了定位，并将其平移至模板中心，再进行特征提取，然后采用基于多层激励函数的量子神经网络对字符进行识别。采用matlab进行仿真实验，结果表明量子神经网络具有较好的识别效率，准确率可达90%以上，抗噪能力强，可以更好的分类。这说明系统的确可以从一定程度上达到提高识别正确率的效果，达到了预期效果。

Character Recognition System Design Based on Image Feature Extraction And QNN

In order to improve the accuracy and reliability of character recognition, combining the image processing technology with the quantum neural network (QNN), the character recognition system based on QNN is studied.Coarse mesh feature method is used to extract the image features. At the same time, in order to enhance the ability of coarse mesh method to resist the change of position,the character image is located and translated to the center of the template before the feature extraction.Then, the quantum neural network based on multilayer excitation function is used to recognize characters.The simulation experiments using MATLAB show that the quantum neural network has better recognition efficiency, and even the accuracy rate can reach more than 90%, and strong noise immunity, and better classification. This shows that the system can indeed improve the correct rate of recognition to a certain extent and achieved the desired results.