为了提高心电信号的正确检测效率,本文在BP神经网络模型基础上,引入多量子能级传递函数,使产生的量子跃迁神经网络具有更有效的学习和泛化能力,克服了经典BP网络在心电图模糊分类方面的局限性,使网络输出“跃变”跳出特征空间不确定区域而“坍缩”到确定区域,提高了心电信号正确检测效率。并且本文为了进一步优化网络,利用具有多分辨特性的小波变换对心电信号进行预处理,减少了网络神经元选取个数,优化网络,提高了收敛速度,实验结果验证了理论的可行性。 In order to improve the correct detection efficiency of ECG signal, this paper introduces multiple quantum level transfer function based on BP neural network model to make the generated quantum transition neural network have more effective learning and generalization ability, Electrocardiogram fuzzy classification limitations, the network output “jump ” jump out of the uncertainty of the region and “collapse ” to determine the area, improve the correct detection of ECG signal efficiency. In order to further optimize the network, wavelet transform with multi-resolution is used to pre-process the ECG signal, which reduces the number of neurons in the network, optimizes the network and improves the convergence speed. The experimental results verify the feasibility of the theory.