Downloads · 30 days
0
raj5517/yolov11s-skin-lesion-isic2018
yolov11s-skin-lesion-isic2018 is a object detection model from raj5517. Use it when you need objects located in an image. The card lists the license as mit.
Fine-tuned YOLOv11s on ISIC 2018 Task 3 for dermoscopic skin lesion detection and classification, with Grad-CAM++ explainability analysis.
Downloads · 30 days
0
Access
Public
Updated Mar 11, 2026
Repo size
76.6 MB
Likes
0
Public
Click a slice to open those files.
.pt76.3 MB · 100%
From the Hugging Face model README
Fine-tuned YOLOv11s on ISIC 2018 Task 3 for dermoscopic skin lesion detection and classification, with Grad-CAM++ explainability analysis.
| File | Description |
|---|---|
best_v2.pt | Best model — 80 epochs, cos_lr (recommended) |
best.pt | v1 baseline — 50 epochs |
results_v2.png | Training curves v2 |
| ID | Class | Full Name | Train Samples |
|---|---|---|---|
| 0 | MEL | Melanoma | 1113 |
| 1 | NV | Melanocytic Nevus | 6705 |
| 2 | BCC | Basal Cell Carcinoma | 514 |
| 3 | AKIEC | Actinic Keratosis / Intraepithelial Carcinoma | 327 |
| 4 | BKL | Benign Keratosis | 1099 |
| 5 | DF | Dermatofibroma | 115 |
| 6 | VASC | Vascular Lesion | 142 |
| Metric | v1 (50 epochs) | v2 (80 epochs + cos_lr) |
|---|---|---|
| [email protected] | 0.551 | 0.603 |
| [email protected]:0.95 | 0.473 | 0.526 |
| Precision | 0.486 | 0.541 |
| Recall | 0.585 | 0.595 |
| Class | [email protected] | Change vs v1 |
|---|---|---|
| MEL | 0.546 | +2.4% |
| NV | 0.956 | +0.7% |
| BCC | 0.556 | +2.5% |
| AKIEC | 0.441 | +14.8% |
| BKL | 0.569 | +0.6% |
| DF | 0.200 | ~flat |
| VASC | 0.850 | +5.9% |
Grad-CAM++ was applied to the backbone (C3k2 layer 8) to visualize which regions the model attends to during inference.
1. Border Ring Detection On well-defined lesions, the model consistently focuses on the lesion perimeter rather than the center. This aligns with clinical dermoscopy criteria where border irregularity is a primary diagnostic indicator — learned without explicit supervision.
2. Multi-focal Pigment Tracking On irregular lesions, the model distributes attention across multiple pigment-dense sub-regions simultaneously, mirroring how dermatologists assess pigment distribution patterns.
These clinically meaningful attention patterns emerged from detection training alone — no segmentation masks, no border annotations, no explicit feature supervision. The model discovered dermoscopy-relevant features autonomously.
In low-confidence cases, Grad-CAM shows the backbone correctly localizes the lesion but detection confidence falls below threshold. This is a known limitation of single-stage detectors on small lesions — the backbone sees it, the detection head doesn't commit.
from ultralytics import YOLO
model = YOLO("raj5517/yolov11s-skin-lesion-isic2018/best_v2.pt")
results = model("dermoscopy_image.jpg")
results[0].show()