Publication detail

Kinect-Supported Dataset Creation for Human Pose Estimation

BEHÚŇ, K. HEROUT, A. PÁLDY, A.

Original Title

Kinect-Supported Dataset Creation for Human Pose Estimation

English Title

Kinect-Supported Dataset Creation for Human Pose Estimation

Type

conference paper

Language

en

Original Abstract

Dataset creation of annotated human poses is very time-consuming. This paper presents an approach for rapid construction of a precisely annotated training dataset for human pose estimation of a sitting subject, intended especially for aeronautic cockpit. We propose to use Kinect as a tool for collecting ground truth to a purely visual dataset (for reasons defined by the application, use of Kinect or similar structured light-based approaches is impossible).  Since Kinect annotation of individual joints might be imprecise at certain moments, manual post-processing of the acquired data is necessary and we propose a scheme for efficient and reliable manual post-annotation. We produced a dataset of 6,322 annotated frames, involving 11 human subjects recorded in various light condition, different clothing, and varying background. Each frame contains one seated person in frontal view with annotation of pose and optical flow data.

English abstract

Dataset creation of annotated human poses is very time-consuming. This paper presents an approach for rapid construction of a precisely annotated training dataset for human pose estimation of a sitting subject, intended especially for aeronautic cockpit. We propose to use Kinect as a tool for collecting ground truth to a purely visual dataset (for reasons defined by the application, use of Kinect or similar structured light-based approaches is impossible).  Since Kinect annotation of individual joints might be imprecise at certain moments, manual post-processing of the acquired data is necessary and we propose a scheme for efficient and reliable manual post-annotation. We produced a dataset of 6,322 annotated frames, involving 11 human subjects recorded in various light condition, different clothing, and varying background. Each frame contains one seated person in frontal view with annotation of pose and optical flow data.

Keywords

Human Pose Estimation, Kinect RGBD, Visual Recognition

RIV year

2014

Released

28.05.2014

Publisher

Comenius University in Bratislava

Location

Smolenice

ISBN

978-80-223-3601-7

Book

Proceedings of Spring Conference on Computer Graphics

Edition

NEUVEDEN

Edition number

NEUVEDEN

Pages from

88

Pages to

95

Pages count

8

Documents

BibTex


@inproceedings{BUT111621,
  author="Kamil {Behúň} and Adam {Herout} and Alexander {Páldy}",
  title="Kinect-Supported Dataset Creation for Human Pose Estimation",
  annote="Dataset creation of annotated human poses is very time-consuming. This paper
presents an approach for rapid construction of a precisely annotated training
dataset for human pose estimation of a sitting subject, intended especially for
aeronautic cockpit. We propose to use Kinect as a tool for collecting ground
truth to a purely visual dataset (for reasons defined by the application, use of
Kinect or similar structured light-based approaches is impossible).  Since Kinect
annotation of individual joints might be imprecise at certain moments, manual
post-processing of the acquired data is necessary and we propose a scheme for
efficient and reliable manual post-annotation. We produced a dataset of 6,322
annotated frames, involving 11 human subjects recorded in various light
condition, different clothing, and varying background. Each frame contains one
seated person in frontal view with annotation of pose and optical flow data.",
  address="Comenius University in Bratislava",
  booktitle="Proceedings of Spring Conference on Computer Graphics",
  chapter="111621",
  doi="10.1145/2643188.2643195",
  edition="NEUVEDEN",
  howpublished="print",
  institution="Comenius University in Bratislava",
  year="2014",
  month="may",
  pages="88--95",
  publisher="Comenius University in Bratislava",
  type="conference paper"
}