Table 4 Classification evaluation results not fine-tuned versions of the model pre-trained on the DTDÂ data set and an endoscopic data set using self-supervison
From: Automated classification of polyps using deep learning architectures and few-shot learning
Model | Classification | Acc | Pre | Rec | F1 |
|---|---|---|---|---|---|
DTD | 1-nn | 68.13 (± 14.21) | 69.48 (± 12.25) | 68.13 (± 14.21) | 68.8 (± 15.2) |
|  | centroid | 67.69 (± 10.81) | 72.55 (± 11.37) | 67.69 (± 10.88) | 70.0 (± 12.6) |
|  | SVM | 68.64 (± 11.83) | 71.93 (± 10.13) | 68.64 (± 11.83) | 70.1 (± 13.7) |
self-sv. | 1-nn | 65.34 (± 10.50) | 65.91 (± 11.43) | 65.34 (± 10.50) | 65.7 (± 12.9) |
|  | centroid | 65.38 (± 15.32) | 67.72 (± 14.71) | 65.38 (± 15.32) | 66.5 (± 15.9) |
|  | SVM | 72.55 (± 13.82) | 73.95 (± 12.61) | 72.55 (± 13.82) | 73.3 (± 14.7) |