Table 3.

Classification performances of different feature representations in for the set of 7 basic emotions. Numbers in bold highlight the maximum classification rates achieved in each column. PO—Positions and orientation, upper and lower body, POU—Positions and orientation, upper body, P—Positions, upper and lower body, O—Orientation, upper and lower body, PU—Positions, upper body, OU—Orientation, upper body.

Features Set	PO		POU		P		O		PU		OU
Error Rate	3	5	3	5	3	5	3	5	3	5	3	5
CNN	56.6	54.8	38.8	38.8	58.1	56.8	33.6	33.0	41.6	38	50.2	49.0
RNN	55.4	55.2	49.2	49.0	59.4	59.4	36.4	33.8	54.6	54.2	34.4	31.8
RNN-LSTM	65.2	59.6	64.6	61.25	69.0	67.0	55.0	54.6	65.8	64.2	54.2	53.8
ResNet20	-	-	-	-	27.8	27.5	-	-	25	23.7	-	-