Desktop Layout

Works in Progress
Colonial Ballroom
Drinks will be served.
Estimation of Racket Grip Vibration from Tennis Video using Neural Network
Kentaro Yoshida, Yuuki Horiuchi, Tomohiro Ichiyama, Seki Inoue, Yasutoshi Makino, and Hiroyuki Shinoda
(University of Tokyo, Japan)
WIP Poster B5
Publisher's Version
Picture (Local)
Supplementary Material
Abstract: In this work, vibrotactile signal felt by a person in a tennis video is automatically estimated from the visual and audio information by using neural network. The system is based on a similar concept to VibVid proposed by the authors. We believe that VibVid system can greatly reduce the effort to describe mathematical models to generate tactile information from diverse videos. In this paper, we try more general and difficult task than the previous research in order to examine the system robustness. We limit the video scene to the back view of a tennis player rallying, but other factors such as locations, player’s clothes, sound environments are arbitrary. We use tennis videos taken in three locations for neural network learning of the relation between the video and measured acceleration of the racket grip. Then we show the grip sensation can be successfully estimated from an unknown video which is taken in a different location from learning. In this case, we plan to examine how to recognize shots more accurately and reproduce high quality tactile sensations.


Time stamp: 2019-03-24T01:30:02+01:00