Testing Reliability of the Computational Age-At-Death Estimation Methods between Five Observers Using Three-Dimensional Image Data of the Pubic Symphysis.

In an effort to standardize data collection and analysis in age estimation, a series of computational methods utilizing high-dimensional image data of the age indicator have recently been proposed as an alternative to subjective visual, trait-to-phase matching techniques. To systematically quantify the reproducibility of such methods, we investigate the intrascan variability and within- and between-observer reliability in initial scan data capturing and editing using 3D laser scans of the Suchey-Brooks pubic symphysis casts and five shape-based computational methods. Our results show that (i) five observers with various training background and experience levels edited the scans consistently for all three trials and the derived shape measures and age estimates were in excellent agreement among observers, and (ii) the computational methods are robust to a measured degree of scan trimming error. This study supports the application of computational methods to 3D laser scanned images for reliable age-at-death estimation, with reduced subjectivity.