Case study
Conquering Skin Cancer
with AI.
Skinmap’s AI solution identifies, monitors, and compares skin lesions from full-body images, helping to deliver faster and more accurate skin cancer screening.
Skinmap’s AI solution identifies, monitors, and compares skin lesions from full-body images, helping to deliver faster and more accurate skin cancer screening.
Skinmap is an AI‑enabled total body photography (TBP) platform developed by Boca Raton-based Triangulate Labs, which helps dermatology professionals detect and track skin changes associated with skin cancer. It delivers fast, clinic‑friendly full‑body scans - about 60 seconds via a mobile app - with real‑time feedback, making TBP more accessible and cost‑effective than traditional systems. The product includes change‑detection features (like a “blink comparator”) to visualize lesion evolution over time, and is rolling out with U.S. dermatology groups through a provider program.
Skinmap’s technical challenge was achieving accurate lesion segmentation across diverse skin types and aligning images in 3D to track subtle lesion changes over time. This required handling subject orientation, inconsistent lighting, shadows, and body hair. Publicly available pretraining data consisted of high‑resolution dermoscopic images, while Skinmap relied on smartphone photos, necessitating pipeline adaptations to improve segmentation performance in real‑world conditions.
The Umaneo team partnered with Skinmap to build a pipeline addressing three core needs: accurate lesion segmentation, precise 3D spatial registration for patient “digital twins,” and longitudinal change‑detection across visits. Leveraging transfer learning, active‑learning with dermatologist review, and streamlined MLOps, the team quickly refined models to prioritize high‑risk lesions and link findings back to source images. Skinmap praised the team’s expertise and speed, noting the AI solution was delivered in nearly half the originally forecasted time, with improved accuracy and workflow integration.
