Top-Down Human Pose Estimation with Depth Images and Domain Adaptation

Nelson Rodrigues, Helena Torres, Bruno Oliveira, João Borges, Sandro Queirós, José Henrique Brito, Jaime Fonseca, Victor Coelho, José Henrique Brito

2019

Abstract

In this paper, a method for estimation of human pose is proposed, making use of ToF (Time of Flight) cameras. For this, a YOLO based object detection method was used, to develop a top-down method. In the first stage, a network was developed to detect people in the image. In the second stage, a network was developed to estimate the joints of each person, using the image result from the first stage. We show that a deep learning network trained from scratch with ToF images yields better results than taking a deep neural network pretrained on RGB data and retraining it with ToF data. We also show that a top-down detector, with a person detector and a joint detector works better than detecting the body joints over the entire image.

Paper Citation

in Harvard Style

Rodrigues N., Torres H., Oliveira B., Borges J., Queirós S., Mendes J., Fonseca J., Coelho V. and Brito J. (2019). Top-Down Human Pose Estimation with Depth Images and Domain Adaptation. In Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2019) - Volume 5: VISAPP; ISBN 978-989-758-354-4, SciTePress, pages 281-288. DOI: 10.5220/0007344602810288

in Bibtex Style

@conference{visapp19,
author={Nelson Rodrigues and Helena Torres and Bruno Oliveira and João Borges and Sandro Queirós and José Henrique Mendes and Jaime Fonseca and Victor Coelho and José Henrique Brito},
title={Top-Down Human Pose Estimation with Depth Images and Domain Adaptation},
booktitle={Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2019) - Volume 5: VISAPP},
year={2019},
pages={281-288},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0007344602810288},
isbn={978-989-758-354-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2019) - Volume 5: VISAPP
TI - Top-Down Human Pose Estimation with Depth Images and Domain Adaptation
SN - 978-989-758-354-4
AU - Rodrigues N.
AU - Torres H.
AU - Oliveira B.
AU - Borges J.
AU - Queirós S.
AU - Mendes J.
AU - Fonseca J.
AU - Coelho V.
AU - Brito J.
PY - 2019
SP - 281
EP - 288
DO - 10.5220/0007344602810288
PB - SciTePress