Table 5.
Comparison of accuracy with existing models.
| S. No | Author(s) | Methodology | Key findings | Accuracy (%) |
|---|---|---|---|---|
| 1 | Turkoglu et al.13 | Pre-trained models + LSTM and SVM classifiers | The LSTM classifier outperformed SVM in accuracy in apple disease detection | 98.20 |
| 2 | Patil & Kumar et al.14 | CNN-LSTM hybrid (Paddy-Fusion) | Outperformed unimodal approaches (e.g., CNN, MLP) | 95.31 |
| 3 | Lamba et al.15 | CNN-LSTM | Achieved varied accuracies for severity classes | 92 |
| 4 | Kukreja et al.18 | LSTM-CNN hybrid model using temporal and spatial data | Effective for determining disease severity | 94.06 |
| 5 | Kaur et al.16 | CNN-LSTM hybrid for Paddy Sheath Rot Disease | Accurate severity classification using temporal and spatial data | 94.84 |
| 6 | Choubey & Dubey19 | CNN + ABi-LSTM with SVM-RFE + ARO for feature selection | High accuracy in plant leaf disease detection | 98.86 |
| 7 | Jiang et al.17 | CNN for leaf image classification | High accuracy in detecting bacterial leaf blight, brown spot, and paddy blast | > 95 |
| 8 | Proposed model | SSDHR classifier | High accuracy | 99.25 |