A new hierarchical method for markerless human pose estimation

Authors:
Yuan Lei;Huawei Pan;Weixia Chen;Chunming Gao
Affiliations:
School of Information Science and Engineering, Hunan University, Changsha, Hunan, China;School of Information Science and Engineering, Hunan University, Changsha, Hunan, China;School of Information Science and Engineering, Hunan University, Changsha, Hunan, China;School of Information Science and Engineering, Hunan University, Changsha, Hunan, China
Venue:
ICVS'13 Proceedings of the 9th international conference on Computer Vision Systems
Year:
2013

Citing 8
Cited 0

A Flexible New Technique for Camera Calibration

IEEE Transactions on Pattern Analysis and Machine Intelligence
Recovering 3D Human Pose from Monocular Images

IEEE Transactions on Pattern Analysis and Machine Intelligence
Markerless tracking of complex human motions from multiple views

Computer Vision and Image Understanding - Special issue on modeling people: Vision-based understanding of a person's shape, appearance, movement, and behaviour
Interacting and Annealing Particle Filters: Mathematics and a Recipe for Applications

Journal of Mathematical Imaging and Vision
HumanEva: Synchronized Video and Motion Capture Dataset and Baseline Algorithm for Evaluation of Articulated Human Motion

International Journal of Computer Vision
Multiview human pose estimation with unconstrained motions

Pattern Recognition Letters
Multi-view 3D Human Pose Estimation in Complex Environment

International Journal of Computer Vision
Loose-limbed People: Estimating 3D Human Pose and Motion Using Non-parametric Belief Propagation

International Journal of Computer Vision

Quantified Score

Hi-index	0.00

Visualization

Abstract

We present a system for markerless human motion capture through a hierarchical method from multiple camera views. In the absence of markers, the task of recovering the human pose is challenging and requires strong image features and robust algorithm. We propose a solution which integrates the 2D posture information and the volumetric reconstruction. Firstly, the model's initia posture is obtained through the method of segmenting silhouette. After that, we track the human pose by using a hierarchical method, which is divided into three steps: head detection, torso prediction and limb matching. In order to gain the robust results, we discard the interior voxel data, use the middle voxel data for motion tracking, and use the surface voxel data for global optimization. The experiment results show that the method is valid and robust.