Uncertainty-Aware Remaining Lifespan Prediction from Images
This addresses the problem of noninvasive health screening for medical research, though it is incremental as it builds on existing foundation models with uncertainty quantification.
The paper tackles predicting remaining lifespan from facial and whole-body images using pretrained vision transformers, achieving state-of-the-art mean absolute errors of 7.41 years on an established dataset and 4.91-4.99 years on new datasets, with calibrated uncertainty estimates.
Predicting mortality-related outcomes from images offers the prospect of accessible, noninvasive, and scalable health screening. We present a method that leverages pretrained vision transformer foundation models to estimate remaining lifespan from facial and whole-body images, alongside robust uncertainty quantification. We show that predictive uncertainty varies systematically with the true remaining lifespan, and that this uncertainty can be effectively modeled by learning a Gaussian distribution for each sample. Our approach achieves state-of-the-art mean absolute error (MAE) of 7.41 years on an established dataset, and further achieves 4.91 and 4.99 years MAE on two new, higher-quality datasets curated and published in this work. Importantly, our models provide calibrated uncertainty estimates, as demonstrated by a bucketed expected calibration error of 0.82 years on the Faces Dataset. While not intended for clinical deployment, these results highlight the potential of extracting medically relevant signals from images. We make all code and datasets available to facilitate further research.