EARS: A System for Geometric and Anthropometric Evaluation of Human Body Scans

Abstract

We present the Enhanced Anthropometric Rating System (EARS), an automated system for evaluating the quality of 3D human body scans. EARS is able to detect and classify both the geometric and anthropometric features of a given mesh and rates its quality. These features and corresponding operations include the roughness of the scanned surface, the fairness of vertex location, area and position of missing body parts, anthropometrically guided segmentation, detection of landmarks, and wrinkles in clothing. The system ranks these features and operations based on their importance as determined by Anthropologists who have specific requirements with respect to understanding the anthropometry of the soldier of the 21st century. The data scans contain more than 100,000 vertices and over 300,000 facets. The system is able to provide real-time feedback on whether the mesh is suitable for downstream applications. The system will be used by the U.S. Army to do statistical studies on their large human body dataset.

Keywords:

• geometry processing
• anthropometric landmarks
• mesh evaluation