Transmission of correct gaze direction in video conferencing using screen-embedded cameras
- PDF / 3,333,586 Bytes
- 18 Pages / 439.642 x 666.49 pts Page_size
- 83 Downloads / 175 Views
Transmission of correct gaze direction in video conferencing using screen-embedded cameras Kazuki Kobayashi1 · Takashi Komuro1 · Keiichiro Kagawa1 · Shoji Kawahito1 Received: 13 December 2019 / Revised: 20 July 2020 / Accepted: 28 August 2020 / © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract In this paper, we propose a new video conferencing system that presents correct gaze directions of a remote user by switching among images obtained from multiple cameras embedded in a screen according to a local user’s position. Our proposed method reproduces a situation like that in which the remote user is in the same space as the local user. The position of the remote user to be displayed on the screen is determined so that the positional relationship between the users is reproduced. The system selects one of the embedded cameras whose viewing direction towards the remote user is the closest to the local user’s viewing direction to the remote user’s image on the screen. As a result of quantitative evaluation, we confirmed that, in comparison with the case using a single camera, the accuracy of gaze estimation was improved by switching among the cameras according to the position of the local user. Keywords Gaze preservation · Teleconferencing · Camera selection
1 Introduction Gaze plays an important role in human face-to-face communication. However, in video conferencing systems using ordinary cameras and displays, there is a problem that the gaze direction is not correctly transmitted since the cameras are typically placed above or below the display. For example, in a video conferencing system in which the camera is fixed above the display, a local user gets the impression that a remote user is always looking down. Therefore, a video conferencing system that is capable of presenting the correct viewing direction is required. To achieve this, video conferencing systems using mechanically movable displays have been developed [9, 11, 13, 16, 17]. These systems transmit the remote user’s viewing direction by changing the direction of the display according to head tracking information for that user or by operating a mouse, for example. However, the systems with movable displays Takashi Komuro
[email protected] 1
Shizuoka University, Johoku 3-5-1, Naka-ku, Hamamatsu, Japan
Multimedia Tools and Applications
suffer from the problem that the devices are expensive and easy to break since they have mechanical parts. Research has also been conducted on telepresence systems using free viewpoint images, allowing images to be generated from arbitrary viewpoint positions [1, 3, 4, 6, 7, 10, 15, 18–20]. These systems can reproduce the viewing direction by acquiring three-dimensional information of the space where the remote user is present and generating images viewed from the viewpoint position of the local user. However, in systems using free viewpoint images, the quality of the synthesized images is limited. There are also systems using a multi-view display and multiple cameras that present
Data Loading...
