Table 2 Malignancy prediction performance where ground truth lesions are manually drawn and annotated by a dermatology trained medical doctor (MK).
From: Lesion identification and malignancy prediction from clinical dermatological images
| Â | AUC | AP |
|---|---|---|
All lesions | 0.790 | 0.775 |
Lesions from smartphone images | 0.789 | 0.775 |
Lesions from dermoscopy images | 0.791 | 0.790 |
