基于脑电信号的情绪特征提取与分类
|
摘要
情绪作为人脑的高级功能,对人们的个性特征和心理健康有很大的影响,利用网上公开的脑电情绪数据库(Deap数据库),根据心理较价和激励唤醒度等级进行情绪划分,对压力和平静两种情绪进行研究分析。在利用db4小波分解与重构算法分解信号的基础上,根据左右脑脑电在产生情绪时脑电信号非对称性的特点,提出一种新的情感特征提取方法,通过计算右侧导联的微分熵值除以左、右对称导联的微分熵之差与右侧导联的微分熵值除以左、右对称导联的微分熵之和,提取出脑电信号的不对称熵特征。利用遗传算法优化的支持向量机对情绪分类识别,平均识别率为88.625%,对比传统特征的分类识别率,利用不对称熵特征的分类识别率有明显提高。
As an advanced function of human brain,emotion has a great impact on people's personality characteristics and mental health. By using the online Deap database,emotions are divided according to psychological valence and arousal level,and the two emotions of stress and calm are studied and analyzed. On the basis of using db4 wavelet decomposition and reconstruction algorithm to decompose the signal,according to the characteristics of the asymmetry of EEG signals in the generation of emotions,a new method of emotional feature extraction is proposed,By dividing the differential entropy of right leads by the difference between left and right symmetrical electrodes,and dividing the differential entropy of right leads by the sum of the differential entropy of left and right symmetrical electrodes,the asymmetric entropy characteristics of EEG signals is extracted. Using the support vector machine optimized by genetic algorithm for emotion classification recognition,the average recognition rate is 88.625%.Comparing with the classification recognition rate of traditional features,the classification recognition rate using the asymmetric entropy feature is significantly improved.
As an advanced function of human brain,emotion has a great impact on people's personality characteristics and mental health. By using the online Deap database,emotions are divided according to psychological valence and arousal level,and the two emotions of stress and calm are studied and analyzed. On the basis of using db4 wavelet decomposition and reconstruction algorithm to decompose the signal,according to the characteristics of the asymmetry of EEG signals in the generation of emotions,a new method of emotional feature extraction is proposed,By dividing the differential entropy of right leads by the difference between left and right symmetrical electrodes,and dividing the differential entropy of right leads by the sum of the differential entropy of left and right symmetrical electrodes,the asymmetric entropy characteristics of EEG signals is extracted. Using the support vector machine optimized by genetic algorithm for emotion classification recognition,the average recognition rate is 88.625%.Comparing with the classification recognition rate of traditional features,the classification recognition rate using the asymmetric entropy feature is significantly improved.
引文
[1]耿雪山,佘青山.基于人工蜂群优化高斯过程的运动想象脑电信号分类[J].传感技术学报,2017,30(3):378-384.
[2]王薇蓉,张雪英,孙颖,等.关于脑电信号的情感识别优化仿真[J].计算机仿真,2018,35(6):426-431.
[3] Duan R N,Zhu J Y,Lu B L. Differential Entropy Feature for EEGBased Emotion Classification//6th Internation IEEE/EMBS Conference on Neural Engineering(NER). SanDiego California,USA:IEEE,2013:81-84.
[4] Krisnandhika B,Faqih A,Pumamasari P D,et al. Emotion Rrecognition System Based on EEG Signals Using Relative Wavelet Energy Features and a Modified Radial Basis Function Neural Networks[C]//International Conference on Consumer Electronics and Devices. London,England,2017:50-54.
[5] Murugappan M,Rizon M,Nagarajan R,et al. Combining Spatial Filtering and Wavelet Transform for Classifying Human Emotions Using EEG Signals[J]. Journal of Medical and Biological Engineering,2011,31(1):45-51.
[6]黄柠檬.基于EEG的情绪识别[D].广州:华南理工大学,2016:48-54.
[7] Koelstra S,Muhl C,Soleymani M,et al. DEAP:A Database for Emotion Analysis; Using Physiological Signals[J]. IEEE Transactions on Affective Computing,2012,3(1):18-31.
[8]陈明.基于脑电信号的情绪识别[D].杭州:杭州电子科技大学,2017.
[9]谢佳利.多尺度熵算法在情感脑电识别中的应用[D].燕山:燕山大学.2016:32-33.
[10]李昕,蔡二娟,田彦秀,等.一种改进特征提取算法及其在情感识别中的应用[J].生物医学工程学杂志,2017,34(4):515-516.
[11]李立.基于脑电信号样本熵的情感识别[D].太原:太原理工大学,2014:26-28.
[12]于亚萍,孙立宁,张锋峰,等.基于小波变换的多特征融合s EMG模式识别[J].传感技术学报,2016,29(04):512-518.
[13]周正杰.情感识别中EEG信号的特征表示研究[D].北京:中央民族大学,2015.
[14]姜媛,林崇德.情绪的脑单侧化研究进展[J].心理与行为研究,2010,8(4):312-318.
[15] Louis A Schmidt,Laurel J Trainor. Frontal Brain Electrical Activity(EEG)Distinguishes Valence and Intensity of Musical Emotions[J]. Cognition and Emotion,2001,15(4):487-500.
[16]杨朝,何明浩,韩俊,等.一种新的支持向量机核函数评估方法[J].雷达科学与技术.2017,15(6):631-634.
[17]曹路,欧阳效源.基于遗传算法的支持向量机的参数优化[J].计算机与数字工程,2016,44(4):575-577.
[18]马静,李星野,徐荣.基于高斯核支持向量机和遗传算法的优化组合研究[J].经济数学,2017,34(1):12-17.
[19] Teodiano Freire Bastos-Filho,Andre Ferreia,Anibal Cotrina Atencio,et al. Evaluation of Feature Extraction Techniques in Emotional State Recognition[C]//IEEE Proceedings of 4thInternational Conference on Intelligent Human Computer Interaction,Kharagpur,India,December. 2012:27-29.
[20]苏建新.基于脑电信号的情绪识别研究[D].南京:南京邮电大学,2015:48-49.
[2]王薇蓉,张雪英,孙颖,等.关于脑电信号的情感识别优化仿真[J].计算机仿真,2018,35(6):426-431.
[3] Duan R N,Zhu J Y,Lu B L. Differential Entropy Feature for EEGBased Emotion Classification//6th Internation IEEE/EMBS Conference on Neural Engineering(NER). SanDiego California,USA:IEEE,2013:81-84.
[4] Krisnandhika B,Faqih A,Pumamasari P D,et al. Emotion Rrecognition System Based on EEG Signals Using Relative Wavelet Energy Features and a Modified Radial Basis Function Neural Networks[C]//International Conference on Consumer Electronics and Devices. London,England,2017:50-54.
[5] Murugappan M,Rizon M,Nagarajan R,et al. Combining Spatial Filtering and Wavelet Transform for Classifying Human Emotions Using EEG Signals[J]. Journal of Medical and Biological Engineering,2011,31(1):45-51.
[6]黄柠檬.基于EEG的情绪识别[D].广州:华南理工大学,2016:48-54.
[7] Koelstra S,Muhl C,Soleymani M,et al. DEAP:A Database for Emotion Analysis; Using Physiological Signals[J]. IEEE Transactions on Affective Computing,2012,3(1):18-31.
[8]陈明.基于脑电信号的情绪识别[D].杭州:杭州电子科技大学,2017.
[9]谢佳利.多尺度熵算法在情感脑电识别中的应用[D].燕山:燕山大学.2016:32-33.
[10]李昕,蔡二娟,田彦秀,等.一种改进特征提取算法及其在情感识别中的应用[J].生物医学工程学杂志,2017,34(4):515-516.
[11]李立.基于脑电信号样本熵的情感识别[D].太原:太原理工大学,2014:26-28.
[12]于亚萍,孙立宁,张锋峰,等.基于小波变换的多特征融合s EMG模式识别[J].传感技术学报,2016,29(04):512-518.
[13]周正杰.情感识别中EEG信号的特征表示研究[D].北京:中央民族大学,2015.
[14]姜媛,林崇德.情绪的脑单侧化研究进展[J].心理与行为研究,2010,8(4):312-318.
[15] Louis A Schmidt,Laurel J Trainor. Frontal Brain Electrical Activity(EEG)Distinguishes Valence and Intensity of Musical Emotions[J]. Cognition and Emotion,2001,15(4):487-500.
[16]杨朝,何明浩,韩俊,等.一种新的支持向量机核函数评估方法[J].雷达科学与技术.2017,15(6):631-634.
[17]曹路,欧阳效源.基于遗传算法的支持向量机的参数优化[J].计算机与数字工程,2016,44(4):575-577.
[18]马静,李星野,徐荣.基于高斯核支持向量机和遗传算法的优化组合研究[J].经济数学,2017,34(1):12-17.
[19] Teodiano Freire Bastos-Filho,Andre Ferreia,Anibal Cotrina Atencio,et al. Evaluation of Feature Extraction Techniques in Emotional State Recognition[C]//IEEE Proceedings of 4thInternational Conference on Intelligent Human Computer Interaction,Kharagpur,India,December. 2012:27-29.
[20]苏建新.基于脑电信号的情绪识别研究[D].南京:南京邮电大学,2015:48-49.