以BP神经网络为工具的短时ECG信号情感分类

Emotional Classification of Short-term ECG Signals Using BP Neural Network

针对目前生理信号情感识别领域采用的生理信号种类太多或使用的生信号长度较长的问题,本文使用BP神经网络对单一、短时ECG信号进行情感识别分类,并对识别时间进行了估计。通过诱发被试喜、怒、哀、惧和平静5种基本情感状态,采集到ECG生理信号,处理后利用神经网络建立模型。实验结果表明,本文方法得到的情感分类的平均识别率为89.14%,且生理信号进行特征提取和识别分类的时间总和小于0.15s,有效地降低了对生理信号种类和窗口长度的依赖。

In response to the problem of too many types of physiological signals or long lengths of biological signals used in the current field of physiological signal emotion recognition,this paper uses BP neural network to classify single and short-term ECG signals for emotion recognition,and estimates the recognition time.By inducing five basic emotional states of joy,anger,sadness,fear,and calmness in the subjects,ECG physiological signals were collected and processed to establish a model using a neural network.The experimental results show that the average recognition rate of emotion classification obtained by the method in this article is 89.14%,and the total time for feature extraction and recognition classification of physiological signals is less than 0.15 seconds,effectively reducing the dependence on physiological signal types and window lengt.