Table 6.
Comparison of the DeepBatch model with the related works using the same rice disease dataset
| Study | Objective | Rice dataset | Best performance (%) | |
|---|---|---|---|---|
| DC | A | |||
| Kanuparthi et al. [62] | Classification of rice diseases by a ternary ensemble of TL models | UCI BLB, BS, and LSm diseases | _ | 96.42 |
| Das and Sengupta [60] | Rice disease detection and classification using image processing and data mining | UCI BLB, BS, and LSm diseases | _ | 95 |
| Patidar et al. [61] | Detection and classification of rice diseases using Deep ResNet | UCI BLB, BS, and LSm diseases | _ | 95.83 |
| Prajapati et al. [41] | Detection and classification of rice diseases using global feature + SVM (Gaussian kernel) | UCI BLB, BS, and LSm diseases | _ | 73.33 |
| DeepBatch (present work) | Segmentation of infected rice regions with a pre-trained model and dilated convolution in encoder Classification of segments by ensemble of TL models | UCI BLB, BS, and LSm diseases | 77.66 | 96.94 |