Seeing beyond BMI: Estimating cardiometabolic risk with smartphone imagery

The thread treated the phone-camera risk estimator less as a breakthrough in BMI replacement than as the next step in a familiar body-composition race. Several readers pointed out that Amazon Halo and Google PhotoScan had already pursued image-based estimates, while smartwatch bioimpedance is an unimpressive baseline because consumer BIA is notoriously crude. The real question is whether the new approach can beat tools with murkier ground truth: even DXA readings can disagree dramatically across machines, so claims of accuracy need more than a favorable comparison against smart scales. The optimistic case is compelling when framed as continuous, near-free monitoring. A phone could combine visual body measurements with wearable activity, resting heart rate, bloodwork, and glucose data, turning occasional clinic tests into a richer longitudinal picture. Some expect major health platforms to absorb that capability, perhaps augmented by phone depth sensors. But enthusiasm kept colliding with the surveillance implications. Readers worried that insurers could turn inferred health risk into premium discrimination, and that handing major platforms systematic images of one’s body is qualitatively creepier than ordinary photo storage. A smaller privacy dispute around importing lab data underscored the same tension: useful health-data interoperability also creates valuable material for fraud, profiling, and abuse.