Automated anthropometric data collection from three-dimensional digital human models

This paper presents the development of an automated anthropometric data-collection system for digital human models obtained from a three-dimensional whole-body scanner. Thirty-seven landmarks were specified for feature recognition and data collection. Both color information and geometric relations were used to develop the algorithms for identifying the landmarks. The proposed landmark identification method has been tested on 105 human models. The recognition rate was over 98%. The experimental results indicate that the proposed method was very efficient and effective in identifying the landmarks. Furthermore, 103 anthropometric dimensions can be extracted from a digital human model. Comparing the data obtained from the proposed system with those obtained by the traditional direct measurement method, the system is fast, accurate, and consistent. Moreover, based on the collected body dimensions, clothing patterns can be directly generated in CAD file format, and textile cutting can be subsequently executed by a CNC laser cuter. An automated tailoring system can thus be developed. The concept of mass customization for the apparel industry can thus be realized.