A Flexible New Technique for Camera Calibration
IEEE Transactions on Pattern Analysis and Machine Intelligence
The Visual Hull Concept for Silhouette-Based Image Understanding
IEEE Transactions on Pattern Analysis and Machine Intelligence
Human Body Model Acquisition and Tracking Using Voxel Data
International Journal of Computer Vision
Inferring 3D Structure with a Statistical Image-Based Shape Model
ICCV '03 Proceedings of the Ninth IEEE International Conference on Computer Vision - Volume 2
M^3: Marker-Free Model Reconstruction and Motion Tracking from 3D Voxel Data
PG '04 Proceedings of the Computer Graphics and Applications, 12th Pacific Conference
Articulated Body Motion Capture by Stochastic Search
International Journal of Computer Vision
Monocular Human Motion Capture with a Mixture of Regressors
CVPR '05 Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) - Workshops - Volume 03
Human Motion De-noising via Greedy Kernel Principal Component Analysis Filtering
ICPR '06 Proceedings of the 18th International Conference on Pattern Recognition - Volume 03
Human pose estimation from monocular image captures
ICME'09 Proceedings of the 2009 IEEE international conference on Multimedia and Expo
Markless Human Motion Capture by Voxel Labeling
ICIG '09 Proceedings of the 2009 Fifth International Conference on Image and Graphics
3D human pose from silhouettes by relevance vector regression
CVPR'04 Proceedings of the 2004 IEEE computer society conference on Computer vision and pattern recognition
Hi-index | 0.00 |
Human body pose estimation is a very important procedure in human posture reconstruction. In order to obtain the initial motion parameter, various model-based methods must use an appropriate posture to track and match a fixed parameter, which is stored in the system. Learning-based methods that rely on a priori probabilities devote much time learning human poses in advance, thus delaying posture initialization. To solve this initialization problem, we propose a new approach in human body pose recovery using ordinary cameras and PCs. We apply a layer-thinning method to obtain the human skeleton points set from voxels, which are obtained in real time using the shape-from-silhouette algorithm based on a lookup table. In the process, the method analyzes the thinned point set, locates human articulation points, and determines the body orientation, using vector analysis according to anthropometric measurements. Through this method, the initial pose can have a variety of free gestures with low-intensity limits. Experiment results show that this approach recovers the initial skeleton pose in a valid and robust manner. In addition, several perturbations in the input data and the human pose recovered rapidly and automatically.