Comparison of body height estimation using bipeds or cylinders

A well-known method for performing body height measurements on images uses a 3D model of the crime scene, projected onto the image. In the 3D model the height of persons on images can be measured using either cylinders or bipeds (3D humanoids). The study compares the performance of taking body height measurements on the basis of both tools. To this, measurements were compared to actual heights in a pilot study that involved images of four perpetrators, one not standing straight. Differences in systematic and random measurement error were studied, showing no significant difference in behavior between the methods. Systematic measurement error is regularly occurring for the biped method as well (two out of four cases), so bipeds are not preferable to cylinders in this respect. Next to this, random measurement errors for the two methods seem comparable. Since the fitting of bipeds is more labor-intensive, there seems to be limited sense in preferring this method to using cylinders to measure body heights in 3D models.