Finding waist levels on 3D scans is challenging because measurement programmes measure bodies non-invasively in a digital environment. However, most body measurement standards require interactions between measurers and subjects, such as bending and touching, to find accurate waist levels. Research shows that the accuracy of commonly used waist substitutes, such as the small of the back and the narrowest from the front view, depends on 3D scans' body shape and size, making the waist substitutes not ideal as generic ways of locating 3D waist levels. This research aimed to study the relationship between 3D waist levels found by domain experts and digital landmark coordinates identified by commercial body measurement programmes to develop regression models that predict waist levels on 3D scans. The developed regression model was tested and found to perform better than the selected waist substitutes. It assists the standardization process of extracting waist-related body measurements from 3D scans.