Convolutional neural network based emotion classification using electrodermal activity signals and time-frequency features
作者：
Highlights：
• EDA based emotion recognition is widely preferred and these signals are highly complex.
• Convolution neural network and short-time Fourier transform are proposed to address this property.
• Thirty-eight features, CNN model and five classifiers are employed along with three learning algorithms.
• Representative and key features are learned using CNN to distinguish various emotions.
• Experiments with public domain database shows their usefulness.
摘要
•EDA based emotion recognition is widely preferred and these signals are highly complex.•Convolution neural network and short-time Fourier transform are proposed to address this property.•Thirty-eight features, CNN model and five classifiers are employed along with three learning algorithms.•Representative and key features are learned using CNN to distinguish various emotions.•Experiments with public domain database shows their usefulness.
论文关键词：Emotion classification,Electrodermal activity,Time–frequency spectrum,Convolutional neural network,Extreme learning machine,Classification
论文评审过程：Received 20 January 2019, Revised 20 April 2020, Accepted 13 May 2020, Available online 30 May 2020, Version of Record 23 June 2020.
论文官网地址：https://doi.org/10.1016/j.eswa.2020.113571
