Article

3D Hand Pose Recognition Over a Wide Area Using Two Omnidirectional Cameras with Field-of-view Division

Authors:

HCI International 2022 - Late Breaking Papers. Multimodality in Advanced Interaction Environments: 24th International Conference on Human-Computer Interaction, HCII 2022, Virtual Event, June 26 – July 1, 2022, Proceedings

Pages 3 - 17

https://doi.org/10.1007/978-3-031-17618-0_1

Published: 26 June 2022 Publication History

Abstract

In this paper, we propose a method for 3D hand pose recognition using two omnidirectional cameras that enables users to perform proximity gesture operations in the entire range in front of a display. In this method, we use a technique of FOV division, which transforms an input omnidirectional camera image into multiple perspective projection images with virtually rotating the camera, in order to avoid distortion in the peripheral area of a perspective projection image. We also introduce two-stage skeleton detection, which uses the results of whole-body skeleton detection for determining the range of hand skeleton detection to reduce false detections. We evaluated the detection rate with and without FOV division. The detection rate with FOV division is higher than that without FOV division, and complex poses can be detected. In addition, the effectiveness of the two-stage skeleton detection was confirmed by comparing the results with and without the two-stage detection.

References

[1]

Kumar P, Verma J, and Prasad S Hand data glove: a wearable real-time device for human-computer interaction Int. J. Adv. Sci. Technol. 2012 43 15-26

Google Scholar

[2]

Pietroszek, K., Tahai, L., Wallace, J.R., Lank, E.: Watchcasting: freehand 3d interaction with off-the-shelf smartwatch. In: Proceedings of 2017 IEEE Symposium on 3D User Interfaces (3DUI), pp. 172–175 (2017)

Google Scholar

[3]

Mäkelä V, James J, Keskinen T, Hakulinen J, and Turunen M “ It’s natural to grab and pull”: retrieving content from large displays using mid-air gestures IEEE Pervas. Comput. 2017 16 3 70-77

Crossref

Google Scholar

[4]

Raj, M., Creem-Regehr, S.H., Rand, K.M., Stefanucci, J.K., Thompson, W.B.: Kinect based 3d object manipulation on a desktop display. In: Proceedings of the ACM Symposium on Applied Perception, pp. 99–102 (2012)

Google Scholar

[5]

Bizzotto N, Costanzo A, Bizzotto L, Regis D, Sandri A, and Magnan B Leap motion gesture control with Osirix in the operating room to control imaging: first experiences during live surgery Surg. Innov. 2014 21 6 655-656

Crossref

Google Scholar

[6]

Fanini, B.: A 3D interface to explore and manipulate multi-scale virtual scenes using the leap motion controller. In: Proceedings of ACHI 2014, the Seventh International Conference on Advances in Computer-Human Interactions, pp. 258–263 (2014)

Google Scholar

[7]

Yam-Viramontes, B.A., Mercado-Ravell, D.: Implementation of a natural user interface to command a drone. In: Proceedings of 2020 International Conference on Unmanned Aircraft Systems (ICUAS), pp. 1139–1144 (2020)

Google Scholar

[8]

Cao Z, Hidalgo G, Simon T, Wei SE, and Sheikh Y Openpose: realtime multi-person 2d pose estimation using part affinity fields IEEE Trans. Pattern Anal. Mach. Intell. 2019 43 1 172-186

Crossref

Google Scholar

Index Terms

3D Hand Pose Recognition Over a Wide Area Using Two Omnidirectional Cameras with Field-of-view Division
1. Computing methodologies
  1. Artificial intelligence
    1. Computer vision
      1. Computer vision problems
      2. Computer vision tasks
        Scene understanding
  2. Computer graphics
2. Human-centered computing
  1. Human computer interaction (HCI)
    1. Interaction devices

Index terms have been assigned to the content through auto-classification.

Recommendations

3D Hand Pointing Recognition over a Wide Area using Two Fisheye Cameras
HCI International 2021 - Late Breaking Papers: Multimodality, eXtended Reality, and Artificial Intelligence
Abstract
In this paper, we propose a user interface for large displays that allows pointing operations from a wide area. Two fisheye cameras installed on both sides of the display are used to capture images of a wide area in front of the display, and the ...
A simplified two-view geometry based external calibration method for omnidirectional and PTZ camera pairs

We present a new external calibration method for an omnidirectional-PTZ camera pair.The method is based on two-view geometry and requires only two scene points.Unlike alternative methods, it does not make very restrictive assumptions on camera ...
Structure from Motion with Wide Circular Field of View Cameras

This paper presents a method for fully automatic and robust estimation of two-view geometry, autocalibration, and 3D metric reconstruction from point correspondences in images taken by cameras with wide circular field of view. We focus on cameras which ...

Comments

Please enable JavaScript to view thecomments powered by Disqus.

Information & Contributors

Information

Published In

Jun 2022

568 pages

ISBN:978-3-031-17617-3

DOI:10.1007/978-3-031-17618-0

Editors:
Masaaki Kurosu
The Open University of Japan, Chiba, Japan
,
Sakae Yamamoto
Tokyo University of Science, Tokyo, Saitama, Japan
,
Hirohiko Mori
Tokyo City University, Tokyo, Japan
,
Dylan D. Schmorrow
Soar Technology Inc., Orlando, FL, USA
,
Cali M. Fidopiastis
Katmai Government Services, Orlando, FL, USA
,
Norbert A. Streitz
Smart Future Initiative, Frankfurt am Main, Germany
,
Shin'ichi Konomi
Kyushu University, Fukuoka, Japan

Publisher

Springer-Verlag

Berlin, Heidelberg

Publication History

Published: 26 June 2022

Author Tags

Qualifiers

Article

Index Terms

Recommendations

3D Hand Pointing Recognition over a Wide Area using Two Fisheye Cameras

A simplified two-view geometry based external calibration method for omnidirectional and PTZ camera pairs

Structure from Motion with Wide Circular Field of View Cameras