Dimension extraction from three dimensional (3D) hand data without prior manual landmarking

摘要

The development of three dimensional (3D) scan technologies makes the anthropometric measurement more efficient and accurate. These markers on human hand are used to facilitate data collection and measurement. An approach of dimension extraction from three dimensional (3D) hand data without prior manual landmarking was proposed in this paper. This is accomplished by developing soft package under the aids of Unigraphics. The first step is to translate the original scanned STL file into the point clouds format. To identify the landmarks from the data cloud point without reference is not an easy task and may involve human errors. So secondly, the program identifies landmarks with the reference of 2D hand image. Subsequently, the program segments the hand into several regions and builds the finger coordinate to bring convenience the dimension calculation. Finally, the program calculates the hand dimensions and save the results for future work. The preliminary results were shown in this paper. The efficiency and robustness of our method deserves more investigation.