TECH Signal 483
Google PhotoScan estimates body composition and insulin resistance from smartphone photos with near-DXA accuracy
↗
Google Research's PhotoScan deep learning framework predicts body composition metrics and insulin resistance from standard 2D smartphone photos, achieving accuracy comparable to DXA scans while requiring no specialized clinical equipment.
Insulin resistance is underdiagnosed and typically requires expensive DXA scans with radiation exposure to assess through body composition. A smartphone-based approach could enable widespread, non-invasive screening for cardiometabolic risk at scale.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
PhotoScan uses a ResNet-50 backbone pre-trained on 35,323 UK Biobank records and fine-tuned on 677 adults with DXA ground truth to estimate body fat percentage, android-to-gynoid ratio, and visceral-to-subcutaneous fat ratio from 2D photos.
The model predicts insulin resistance (HOMA-IR > 2.9) with near-DXA accuracy, outperforming smartwatch BIA sensors on body fat percentage and providing A/G and V/S ratios that BIA cannot measure.
PhotoScan is described as investigational and was validated in clinical research settings, not yet in general clinical deployment.
THE CLUSTER