Semisupervised Deep Features of Time-Frequency Maps for Multimodal Emotion Recognition

Table 1

Summary of multimodal emotion recognition from biological signals.

References	Modality	Dataset	Domain analysis	Fusion/classification
[7]	EEG	DEAP	Time domain	After feature extraction from each modal/decision tree
	PPS	MAHNOB-HCI
[8]	EEG	DEAP	Time-frequency domain (MSST) for EEG	After feature extraction from each modal/CNN
	PPS	MAHNOB-HCI	Time domain for PPS
	Video		Time and frequency domains for video
[9]	EEG	DEAP	Time domain	Before feature extraction/ensemble CNN
	PPS
[10]	EEG	DEAP	Time domain	After feature extraction from each modal/MLP
	PPS
[11]	EEG	DEAP	Time and frequency domains for different modalities	After feature extraction from each modal/CNN
	PPS
[12]	PPS	DEAP	Time-frequency domain	After computing the CWT of each modal/CNN
[17]	EEG	DEAP	Time domain	After feature extraction from each modal by CNN/MLP
	Video	MAHNOB-HCI
[19]	EEG	DEAP	Time domain	After feature extraction from each modal by CNN/Decision tree
	GSR	LUMED-2
	Video
[20]	EEG	DEAP	Time domain	After feature extraction from each modal by 3D-CNN/MLP
	Video
[22]	EEG	Private	Frequency domain	After feature extraction from each modal/CNN
	Audio