Table 3 Performance Comparison of Various Models for Human Activity Classification.
From: Bi-directional ConvLSTM networks for early recognition of human activities and action prediction
Model | Accuracy Values(%) | Mean Accuracy (%) | Precision (%) | Recall (%) | F1-Score (%) | Standard Deviation |
|---|---|---|---|---|---|---|
CNN | [82.00, 82.50, 82.40, 82.30, 82.29] | 82.29 | 82.00 | 81.20 | 82.10 | 0.22 |
InceptionV3 | [86.20, 86.60, 86.40, 86.70, 86.53] | 86.49 | 86.20 | 86.30 | 86.25 | 0.22 |
VGG19 | [79.50, 79.80, 79.70, 79.60, 79.74] | 79.67 | 79.50 | 79.60 | 79.55 | 0.12 |
InceptionV3-LSTM | [83.40, 83.60, 83.50, 83.55, 83.54] | 83.52 | 83.40 | 83.50 | 83.45 | 0.08 |
ConvLSTM | [84.40, 84.60, 84.55, 84.50, 84.53] | 84.52 | 84.40 | 84.50 | 84.45 | 0.08 |
Bi-ConvLSTM | [85.70, 85.80, 85.75, 85.85, 85.78] | 85.77 | 85.70 | 85.75 | 85.72 | 0.06 |
Proposed Bi-ConvLSTM | [89.40, 89.60, 89.50, 89.55, 89.54] | 89.52 | 89.40 | 89.50 | 89.45 | 0.08 |
