Estimation of Human Body Shape in Motion with Wide Clothing

European Conference on Computer Vision, 2016
Jinlong Yang, Jean-Sébastien Franco, Franck Hétroy-Wheeler, and Stefanie Wuhrer
Abstract Paper Video Download dataset Download result













Abstract





Estimating 3D human body shape in motion from a sequence of unstructured oriented 3D point clouds is important for many applications. We propose the first automatic method to solve this problem that works in the presence of loose clothing. The problem is formulated as an optimization problem that solves for identity and posture parameters in a shape space capturing likely body shape variations. The automation is achieved by leveraging a recent robust pose detection method  (Zuffi and Black 2015). To account for clothing, we take advantage of motion cues by encouraging the estimated body shape to be inside the observations. The method is evaluated on a new benchmark containing different subjects, motions, and clothing styles that allows to quantitatively measure the accuracy of body shape estimates. Furthermore, we compare our results to existing methods that require manual input and demonstrate that results of similar visual quality can be obtained.

Dataset & result

To download the whole dataset, please click here.
To download the whole result, please click here.
README.
Last updated on 29 Sep, 2016. Fixed some index bugs in dataset thanks for Chao Zhang.

Clothing type
Tight
Layered
Wide
Motion
Walk
Walk
Spin
Knees-up
Walk
Spin
Knees-up
Sub #1
Dataset Dataset Dataset Dataset Dataset Dataset Dataset
Result Result Result Result Result Result Result
Sub #2
Dataset Dataset Dataset Dataset Dataset Dataset Dataset
Result Result Result Result Result Result Result
Sub #3
Dataset Dataset Dataset Dataset Dataset Dataset Dataset
Result Result Result Result Result Result Result
Sub #4
Dataset Dataset Dataset Dataset Dataset Dataset Dataset
Result Result Result Result Result Result Result
Sub #5
Dataset Dataset Dataset Dataset Dataset Dataset Dataset
Result Result Result Result Result Failure Result
Sub #6
Dataset Dataset Dataset Dataset Dataset Dataset Dataset
Result Result Result Result Result Result Result
*  The dataset was acquired using the Kinovis platform: http://kinovis.inrialpes.fr/
**This project is funded by France National Research grant ANR-14-CE24-0030 ACHMOV.
Dataset license
This dataset aims to evaluate algorithms that estimate human body shape in the presence of loose clothing. The data is made available as is without any warranty; without even the implied warranty of fitness for a particular purpose. The redistribution of the data is not permitted. The data are provided for non-commercial research purposes only.

If you use this data or part of it in a publication, you agree to cite:

Jinlong Yang, Jean-Sébastien Franco, Franck Hétroy-Wheeler, Stefanie Wuhrer.
Estimation of Human Body Shape in Motion with Wide Clothing.
European Conference on Computer Vision, 2016.