Table 5 summarizes the quantitative comparison across accuracy, precision, recall, F1-score, and inference latency.
Model | Accuracy (%) | Precision (%) | Recall (%) | F1-score (%) | Inference latency (ms) |
|---|---|---|---|---|---|
CNN (3 Conv + 1 Dense) | 90.1 | 89.3 | 89.8 | 89.5 | 134 |
LSTM (2 layers) | 91.2 | 90.7 | 91.0 | 90.8 | 156 |
DenseNet (40 Layers) | 90.8 | 91.0 | 90.1 | 90.2 | 128 |
Proposed CNN-LSTM (hybrid) | 91.9 | 91.6 | 91.4 | 91.5 | 118 |
