Skip to main content
EURASIP Journal on Advances in Signal Processing

Table 4 Integrated results on RAF-DB dataset

From: Research on real-time teachers’ facial expression recognition based on YOLOv5 and attention mechanisms

Method

Model Size

Parameters

Accuracy (%)

mAP@0.5 (%)

Time (ms)

FASA

92.201 M

47.295 M

75.1

82.1

20

FA + CBSA + CSPA + SA

112.400 M

55.735 M

75.3

82.4

37

CSPA + SA

96.900 M

51.103 M

75.8

82.0

29

FA + CSPA

96.900 M

51.103 M

76.5

82.7

29

CBSA + CSPA

105.501 M

52.310 M

76.8

83.0

32

  1. Bold meant the best performance of tampering detection in the same experimental setting
Back to article page