Automatic Human Body Feature Extraction and Size Measurement by Random Forest Regression Analysis of Geodesics Distance
PubDate: May 2019
Teams: Capital Normal University； Beihang University
Writers: Xiaohui Tan; Xiaoyu Peng; Liwen Liu; Qing Xia
Abstract
It is a pervasive problem to obtain the human body size without contacting for apparel application. In this paper, a new approach is proposed which can be applied to calculate human body size such as shoulder width, bust, hips and waist girth with single depth camera. First, single depth camera as the 3D model acquisition device was used to get the 3D human body model. Then an automatic extraction method of focal features on 3D human body via random forest regression analysis of geodesics distance is used to extract the predefined feature points and lines. Finally, the individual human body size is calculated according to the feature points and lines. The method is an automatic data-driven way. The scale-invariant heat kernel signature is exploited to serve as feature proximity. So it is insensitive to postures and different shapes of 3D human body. These main advantages of our method lead to a robust and accurate feature extraction technique and size measurement for 3D human bodies in various postures and shapes. The experiment results show that the average relative error of feature points extraction is 0.0617 cm. The average relative errors of shoulder width and girth are 1.332 cm and 0.7635 cm, respectively. Overall, the algorithm has a better detection effect for 3D human body size, and it is stable with better robustness.
