Table 2 Accuracy of classifiers on binary and multi-class data-set.
From: Towards robust diagnosis of COVID-19 using vision self-attention transformer
Refs. | Dataset | Approach | Accuracy (%) | AUROC (%) |
|---|---|---|---|---|
BIMCV, NIH | ViT | 86.9 | 0.92 | |
Brazilian | Machine learning | 87.66 | 0.9056 | |
Brazilian | Voting based | 87.66 | 0.906 | |
Brazilian | xDNN | 97.386 | 0.9736 | |
COVID CT scans | CNN & ConvLSTM | 99 | – | |
Hust19 | Deep learning | Â | 0.9946 | |
COVID-19 CT Dataset | Transfer learning | 83.6 | 0.946 | |
Own Dataset | DeCovNet | 90.16 | 0.9596 | |
Proposed approach | Brazilian | Vision transformer | 98 | 0.996 |
Proposed approach | Hust19 | Vision transformer | 99.7 | 0.997 |
