BROWSE
Detail
Conference Paper
A Method of Estimating the Object of Interest from 3D Object and User’s Gesture in VR
Cited 0 time in 
Share 
Share
- Authors
- Sungjin Hong, Heesook Shin, ChoRong Yu, Seong Min Baek, Youn-Hee Gil
- Issue Date
- 2022-11
- Citation
- Symposium on Virtual Reality Software and Technology (VRST) 2022, pp.1-2
- Language
- English
- Type
- Conference Paper
- Abstract
- In VR, gaze information is useful for directly or indirectly analyzing a user’s interest. However, there are inconveniences in using the eye tracking in the VR device. To overcome the drawback, we propose a method of estimating an object of interest from user’s gesture instead of eye tracking. LightGBM model is trained by using distance and angle-based features that are extracted from 3d information of the object and the position and rotation of the VR device. We compared accuracy of each feature for VR device combinations and found out that it is more efficient to use all devices instead of individual devices and to use angle-based feature instead of distance-based feature with accuracy of 79.36%.
- KSP Keywords
- 3D information, 3D object, Angle-based, Distance-based, Gaze information, eye tracking