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)
SA
91.827 M
46.853 M
74.9
82.1
18
FA
75.3
82.3
CBSA
92.490 M
47.884 M
75.7
82.5
22
CSPA
94.601 M
49.974 M
77.1
83.4
25