A global team of researchers are using a new image collection to train algorithms that can detect skin cancer earlier and extend diagnosis beyond clinical settings. The system includes nearly 17,000 high-resolution photos of skin regions from 100 people, each showing a small section of skin with all visible moles and spots outlined, plus details such as body location, age group, and sun damage level. Unlike older datasets, these images capture surrounding skin and often include multiple lesions, giving algorithms the broader context they need to work reliably in real-world environments.
Image Credits: Hush Naidoo Jade Photography
