Automatic Human Measurement Extraction Algorithms for the Apparel Industry

의복제작을 위한 인체 자동 측정 알고리즘에 관한 연구

A faster and more accurate method of body measurement can benefit made-to-measure operations in the apparel industry .Recently, the development of 3D scanner makes it easy to get 3D human body data. It is possible to use various accurate body measurement data and internet-based made-to-measure operations. In this study we propose computer assisted human boyd land-mark identification algorithms and measurement extraction methods from 3D human scan data. And such database can be used to develop better fitting models for improved standard sizing of garments. When working with computer images, it is impossible to detect landmarks of the surface of a human body, so alternative mathematical procedures must be defined. The procedures we used are geometrical, statistical and practical methods. The data used in this study are from surface coordinate data file of women's body measured by a 3D whole body scanner. The procedure and results were as follows; 1. Review the traditional landmarks and measuring methods and redefine them. 2. Develop automatic identification of human body direction and segmentation. 3. Develop landmark algorithm for the automatic measuring method, and apply it with c++ programming. 4. Compare the results of developed landmark points to those obtained by traditional method.