Dataset · marmal88
Dermatoscopic Skin Lesion Images
marmal88/skin_cancer
The HAM10000 collection of 10,000+ dermatoscopic images for skin lesion classification and segmentation.
What's inside
Contains dermatoscopic images of common pigmented skin lesions from various sources, along with English text data. The dataset has between 10k and 100k records (known as HAM10000 from the eponymous research paper and Kaggle). The data is intended for image classification and segmentation tasks.
What you can do with it
- Train a model to classify types of pigmented skin lesions from images
- Train a segmentation model to isolate lesion boundaries in photos
- Test automated dermatological diagnostic algorithms
- Tasks
- Image classification, Image segmentation
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 2.1K
- Likes
- 47
Models trained on it
- Kuldeepmishra3/vit-large-skin-cancer-ham10000Image classification 237
- efecelik/medgemma-dermatology-loraImage text → text 23
- Ankitdajhgf565/vit-large-skin-cancer-ham10000Image classification 23
- hssling/derm-analyzer-adapter 14
- mufasabrownie/gemma-4-E2B-it-ham10000-loraImage → text 8
- ethicalabs/SkinCancerViTImage classification 0
- fawo/eva02-small-melanoma-classifierImage classification 0
- andy279/convnext-tiny-ham10000-lesion-disjointImage classification 0
How to load it
from datasets import load_dataset
ds = load_dataset("marmal88/skin_cancer")