Body Structure and Anthropometry

With rapid advances in multimedia and network technology, the development of detailed simulation human models have become an important research topic. This research series centers on 3D human body scanning data, advancing automatic feature extraction and measurement of the human body and head.
The study includes:

  1. Automatic Feature Extraction and Anthropometry from Marker-less Human Body Scanning－
    This research proposes a novel computerized anthropometric system for marker-less body scanning. The descriptions of human body features, mostly defined in ASTM and ISO, are interpreted as logical mathematical definitions. Based on these significant definitions, image processing and computational geometry techniques are employed to automatically identify body features from the torso cloud points. With the body surface divided into a series of rectangular patches using the feature lines as boundaries, we reconstructed the body with a minimal number of triangles while retaining the essential shape. In order to obtain an individual’s dimensions quickly and comprehensively, three types of computerized measurement methods are developed on the mesh model. The proposed measuring methods apply to most traditional manual measurements. It aims to present a fast, reliable, and unambiguous method for obtaining human body measurements for use in the garment industry.
  2. Craniofacial structure scalability with features retain－
    This study presents a method for objective and automatic feature identification through searching for geometric feature points and feature lines for the human head which can simplify mesh numbers without losing feature clarity. This study also features a system of automatic craniofacial feature extraction based on MPEG-4 geometric definitions, a method for head skew correction to improve the recognition rate, and integration of computer tomography images for ear reconstruction to resolve ear recognition difficulties. The level of grid detail can be determined according to the requirements of a particular application to create a fully-featured scalable craniofacial model for use in multimedia 3D image transmission, computer animations and other applications.