The NRC Publications Archive is now operational; however, not all the features of the site are available at this time.
The following features remain unavailable:
- Viewing/Downloading of full text publications
NRC is currently working to restore these features and we will update this notice as these features become available. Thank you for your patience.
Human body shape prediction and analysis using predictive clustering tree
; Xi, Pengcheng
; Guo, Hongyu
NRC Institute for Information Technology; National Research Council Canada
International Conference on 3D Imaging, Modeling, Processing, Visualization and Transmission (3DIMPVT 2011): May 16-19, 2011, Hangzhou, China
2011 International Conference on 3D Imaging, Modeling, Processing, Visualization and Transmission (3DIMPVT)
Predictive modeling; digital human modeling; predictive clustering tree; demographic attributes
3D Imaging, Modeling and Visualization; Imagerie 3D, modélisation et visualisation
Digital Human Modeling; Modélisation humaine numérique
Visual Information Technology; Technologie de l'information visuelle
Predictive modeling aims at constructing models that predict a target property of an object based on its descriptions. In digital human modeling, it can be applied to predicting human body shape from images, measurements, or descriptive features. While images and measurements can be converted to numerical values, it is difficult to assign numerical values to descriptive features and therefore regression based methods cannot be applied. In this work, we propose to use Predictive Clustering Trees (PCT) to predict human body shapes from demographic information. We build PCTs using a dataset of demographic attributes and body shape descriptors. We demonstrate empirically that the PCT-based method has similar predicting power as the numerical approaches using body measurements. The PCTs also reveal interesting structures of the training dataset and provide interpretations of the body shape variations from the perspective of the demographic attributes.