Over 2 million adults die each year due to illnesses associated with excess adiposity. Need for non-invasive, accessible measurement of adiposity has led to reliance on waist circumference and BMI. We asked if fat mass and distribution can be accurately measured using 3D optical (3DO) scanning, which many fitness centers have. We estimated fat mass index (FMI), visceral adipose tissue mass (VAT), and created pseudo-DXA fat visualizations from 3DO scans.


From the Shape Up! Adults cohort study, 176 participants’ (72 men) data was available, including Hologic Horizon/A whole-body DXA and Fit3D Proscanner 3DO scans. Like image types were spatially registered using 105 and 75 fiducial points for DXA and 3DO scans, respectively. Statistical appearance and shape modeling were then performed on each image type. The sex-specific population variances for shape (3DO) and appearance (DXA) were captured as principal components (PC) resulting in 4 PC models: PCDXA_men, PCDXA_women, PC3DO_men, PC3DO_women. LASSO regression was used to predict FMI and VAT from PC3DO modes. Stepwise linear regression was used to predict PCDXA modes from PC3DO modes. The predicted PCDXA values of each participant were then inverted to create a pseudo-DXA fat image.


It took 32 PCDXA_men and 38 PCDXA_women modes to describe 95% of the fat variance. Similarly, 10 PC3DOs described 95% of the shape variances. PC3DO accurately estimated FMI and VAT with R^2[RMSE] values of 0.83[1.14 kg/m^2] and 0.59[0.17 kg] for men and 0.94[0.93 kg/m^2] and 0.73[0.15 kg] for women. The pixel-by-pixel differences in fat mass between actual and predicted DXA values had no mean bias and RMSE values of 0.013 g for men and 0.015 g for women.


FMI and VAT can be accurately estimated from 3DO scans in both men and women. Furthermore, DXA fat distribution can be estimated and visualized exclusively from 3DO scans. To our knowledge, this is the first time detailed fat distributions have been derived from 3DO scans.