【摘要】 本文设计了人工神经网络识别带不同噪声信号的英文字母。结果表明:噪声信号的平均水平低于0.2时,神经网络模型能够完全正确识别26个字母;噪声水平在0.2时,神经网络模型开始出现错误,随着噪声水平的增加,神经网络模型的识别率逐渐下降。更多还原

【Abstract】 In this paper,the author used artificial neural network to recognize English letters with noise.The results showed that with the increase of noise levels,the performance of the ANN model decreased: when the noise level was less than 0.2,the letters could be recognized accurately,when the noise level was 0.2,there were some errors in the recognition,then,with the increase of noise levels,there were more and more errors.更多还原