Hybrid Deep Learning Approaches for sEMG Signal-Based Lower Limb Activity Recognition

Table 2

Performance parameters for healthy subjects (in %).

Approach/Parameters	CNN	LSTM	GRU	CNN-LSTM	CNN-GRU
Accuracy	98.88	95.38	98.32	99.02	99.86
Precision	98.49	93.81	97.49	99.20	99.88
Recall/sensitivity	98.13	93.86	97.53	97.88	99.70
Specificity	99.45	97.73	99.20	99.46	99.92
F-score	98.31	93.82	97.51	98.50	99.79