Table 5.
Summary of deep learning methods with image modalities, datasets, and classification results
| Refs | Years | Methods | Datasets | Image Modalities | Results |
|---|---|---|---|---|---|
| [335] | 2020 | (ResNet50, InceptionV3), (SVM,KNN,LDA) | Videos frames | VCE |
97.20% Sen, 95.63% Spe 95.94% Acc |
| [336] | 2020 | SegNet (encoder-decoder) | NBI images | Endoscopy |
98.04% Sen, 95.03% Spe |
| [337] | 2020 | Deep CNN, SVM, Fuzzy | Endoscopy frames | WCE |
82.71% Sen 87.03% Spe |
| [338] | 2019 | (DWT,PSO),(KNN,PNN,DT,SVM) | SB2,SB3 | VCE |
88.43% Sen, 84.60% Spe, 86.45% Acc |
| [87] | 2018 | R-CNN | Videos frames | VCE |
97.30% Sen, 98.00% Spe |
| [339] | 2018 | GAN | CVC | WVE |
74.00% Sen, 94.00% Spe, 91.00%Acc |
| [340] | 2018 | Transfer Learning | Kvasir | WVE | 83.00% Acc |
| [341] | 2018 | GANs | GIANA | VCE |
88.00%Sen, 99.90% Spe, 99.00% Acc |
| [342] | 2017 | CNN | Kvasir | WVE |
75.00% Sen 75.00% Spe |