Stereo Pictorial Structure for 2D Articulated Human Pose Estimation

Manuel I. López-Quintero, Manuel J. Marín-Jiménez, Rafael Muñoz-Salinas, Francisco J. Madrid-Cuevas, Rafael Medina-Carnicer

Abstract

We propose a new technique to automatically detect and estimate the 2D pose of humans in stereo image pairs from realistic stereo videos that can be found in the Internet. To address this task, we propose a novel pictorial structure model to exploit the stereo information included in such stereo image pairs: the Stereo Pictorial Structure (SPS).

To validate our proposed model, we contribute a new annotated dataset of stereo image pairs, the Stereo Human Pose Estimation Dataset (SHPED), obtained from YouTube stereoscopic video sequences, depicting people in challenging poses and diverse indoor and outdoor scenarios.

The experimental results on SHPED indicates that SPS improves on state-of-the-art monocular models thanks to the appropriate use of the stereo information.

Paper

If you find Stereo Pictorial Structure for 2D Articulated Human Pose Estimation useful for your work or you use our software, please cite our paper:

Manuel I. López-Quintero, Manuel J. Marín-Jiménez, Rafael Muñoz-Salinas, Francisco J. Madrid-Cuevas, Rafael Medina-Carnicer,
Stereo Pictorial Structure for 2D Articulated Human Pose Estimation,
Machine Vision and Applications, vol. 27, no. 2, pp. 157–174, 2015.

[PDF]   [DOI]   [BibTeX]

Code

Filename Description Size
sps_v01.zip Stereo pictorial structure software, samples, and third party software 36,1 MB

Dataset

We provide a dataset of stereo image pairs suited for stereo human pose estimation of upper-body people: The Stereo Human Pose Estimation Dataset (SHPED). The description and the download of our dataset can be found here.

Acknowledgements

This work was partially supported by the Research Projects TIN2012-32952 and BROCA, both financed by the Spanish Ministry of Science and Technology and the European Regional Development Fund (FEDER).