Revolutionizing Pathology with UNI
| Type | research |
|---|---|
| Area | Medical |
| Published(YearMonth) | 2403 |
| Source | https://www.nature.com/articles/s41591-024-02857-3 |
| Tag | newsletter |
| Checkbox | |
| Date(of entry) |
A recent publication in Nature Medicine introduces UNI, a general-purpose self-supervised model for computational pathology. UNI, pre-trained with over 100 million images from diagnostic slides, showcases remarkable performance in diagnosing and subtyping 108 cancer types and other diverse pathology tasks. Its robustness and adaptability set a new standard for AI in pathology, promising to enhance diagnostic accuracy and efficiency across multiple tissue types.