Google has taught smartphones to determine body composition from photos – with accuracy approaching X-rays

Google talks about AI-based technology photo scanwhich allows you to determine body composition using a regular camera on your smartphone and provides results that are close to X-ray accuracy.

Last week, Google launched its new Pixel 11 smartphones and Pixel Watch 5 smartwatches, and one of the most important innovations was the Health Guardian toolkit with insulin resistance trending functionality. This feature analyzes the body’s response to blood sugar surges on a monthly basis – important for weight management and controlling the development of type 2 diabetes. An important indicator of insulin resistance and metabolic health is analysis of body composition, that is, the proportions of different types of fat, skeletal muscle, and bone. Body composition analysis includes DXA X-ray scans and some home devices. Samsung Galaxy Watch Ultra 2 features Bioelectrical Impedance Analysis (BIA): it passes a weak electric current through the body, which passes through fat, muscle and bone in different ways, allowing you to estimate their ratios.

As an alternative, Google proposed PhotoScan technology, which only meets the needs of cameras on smartphones and artificial intelligence models that analyze camera images. The search giant developed its PhotoScan technology by matching smartphone images of a person with DXA medical scans and other data about the person, then processing it all using artificial intelligence. The model trained in this way achieves “High DXA compliance” Its accuracy is similar to DXA’s assessment of insulin resistance in assessing body fat percentage and other parameters. This technology is not yet on the market, but “Demonstrates the feasibility of using smartphones to assess body composition”Google said. This solution may appear on Pixel smartphones.