BioinformaticsbioRxivPreprint
Evaluating performance bias in face-to-BMI vision transformer models across diverse human populations
J C Hoffman, Michael Gurven, Hillard S Kaplan…·Thomas S. Kraft
This study evaluates how face-to-BMI Vision Transformer models perform across diverse human populations, finding that training on population-specific data improves accuracy, while global training data enhances performance in underrepresented groups.