Advanced Motion Prediction for Virtual Reality Gaming: a CNN-Based Approach

Abstract
A novel motion prediction model (MPM) for virtual reality (VR) video games was developed, consisting of a motion recognition model (MRM) and a next movement prediction model (NMPM), both using convolutional neural networks (CNNs). Motion capture was performed with HTC Vive Pro and Meta Quest 2. Two custom datasets were created to train the MRM and NMPM. Our method achieved a top-1 accuracy of 77% and a top-2 accuracy of 90%, even with motion data sequences sharing similar initial stages but diverging in subsequent movements.
Description

CCS Concepts: Computing methodologies → Virtual reality; Neural networks; Hardware → Sensor devices and platforms

        
@inproceedings{
10.2312:egve.20241379
, booktitle = {
ICAT-EGVE 2024 - International Conference on Artificial Reality and Telexistence and Eurographics Symposium on Virtual Environments - Posters and Demos
}, editor = {
Tanabe, Takeshi
and
Yem, Vibol
}, title = {{
Advanced Motion Prediction for Virtual Reality Gaming: a CNN-Based Approach
}}, author = {
Jegierski, Hubert
and
Jegierski, Maciej
and
Igras-Cybulska, Magdalena
and
Węgrzyn, Paweł
and
Łapczyński, Adrian
and
Babiuch, Paweł
and
Płaza, Mirosław
and
Pięta, Paweł
and
Łukawski, Grzegorz
and
Deniziak, Stanisław
and
Opałka, Jacek
and
Jasiński, Artur
}, year = {
2024
}, publisher = {
The Eurographics Association
}, ISSN = {
1727-530X
}, ISBN = {
978-3-03868-246-2
}, DOI = {
10.2312/egve.20241379
} }
Citation