Updated
Updated · Android Authority · Aug 19
Google Research Builds PhotoScan AI to Estimate Body Fat, Targeting 73.1% of Overweight U.S. Adults
Updated
Updated · Android Authority · Aug 19

Google Research Builds PhotoScan AI to Estimate Body Fat, Targeting 73.1% of Overweight U.S. Adults

2 articles · Updated · Android Authority · Aug 19

Summary

  • PhotoScan uses a series of smartphone body images to estimate body fat more accurately than wearable bioelectrical impedance sensors, according to Google Research.
  • DXA X-ray data, MRI body data and a smartphone-photo fine-tuning set trained the model, which also estimated android-to-gynoid and visceral-to-subcutaneous fat ratios that wearables cannot measure.
  • Google said those image-based results could also help predict insulin resistance, extending a health push that already includes insulin-resistance tracking planned for Pixel Watch features.
  • No product plans were announced, and adoption may hinge on whether users will share sensitive full-body photos despite privacy concerns around the tool.

Insights

If Google's AI can accurately calculate visceral fat from a photo, are you willing to trade body privacy for health insights?
Will predicting insulin resistance through a simple camera lens fundamentally change how everyday people screen for preventative health?