Table 11.
Summary of ways for reducing or removing overfitting.
| S. No. | Year and Reference | Technique Utilized from Reducing Overfitting |
|---|---|---|
| 1. | 2016 [8] | Distortion was introduced |
| 2. | 2016 [9] | Rotation and flipping (data augmentation) |
| 3. | 2017 [41] | Dropout layers and pooling layers |
| 4. | 2017 [11] | Extensive data augmentation |
| 5. | 2017 [42] | Image processing, response-normalizing layers, swapping out some fully connected layers for convolution layers |
| 6. | 2018 [13] | Data augmentation |
| 7. | 2019 [16] | Distorted images were added |
| 8. | 2019 [62] | Dropout value—0.25 |
| 9. | 2019 [64] | ReLU activation function, data augmentation approaches |
| 10. | 2019 [19] | Early stopping strategy |
| 11. | 2020 [48] | Data augmentation |
| 12. | 2020 [20] | Early stop mechanism, data augmentation techniques, and dropout |
| 13. | 2020 [68] | Data augmentation |
| 14. | 2021 [25] | Data augmentation |
| 15. | 2021 [28] | Data augmentation |
| 16. | 2021 [73] | Early halting |
| 17. | 2021 [53] | 1 × 1 convolution |
| 18. | 2021 [71] | Forward propagation |
| 19. | 2021 [30] | Global average pooling |
| 20. | 2022 [34] | Data augmentation |
| 21. | 2022 [56] | Data augmentation |