Wound Classification on Curated Wound Image Dataset (Medetec, AZH, clinical partnerships) (test)
99.9AccuracyWoundNet-Ensemble
Evaluation results
| Method | Links | |
|---|---|---|
| WoundNet-EnsembleKey Approach=ResNet-50 + DINOv2 + Swin-T Ensemble2025.12 | 99.9 | |
| Multi-modal DNNKey Approach=Image + Body Location Data2025.12 | 97.12 | |
| Swin-TinyKey Approach=Swin Transformer2025.12 | 96.2 | |
| ViT-B/16Key Approach=Vision Transformer2025.12 | 95.8 | |
| EfficientNet-B4Key Approach=CNN (noted top performer)2025.12 | 92.3 | |
| Rostami et al.Key Approach=Ensemble Deep CNN (Patch + Image)2025.12 | 91.9 |