FrictGAN: Frictional Signal Generation from Fabric Texture Images using Generative Adversarial Network

Shaoyu Cai; Yuki Ban; Takuji Narumi; Kening Zhu

doi:10.2312/egve.20201254

FrictGAN: Frictional Signal Generation from Fabric Texture Images using Generative Adversarial Network

Shaoyu Cai, Yuki Ban, Takuji Narumi, Kening Zhu^*

^*Corresponding author for this work

School of Creative Media

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

Abstract

The electrostatic tactile display could render the tactile feeling of different haptic texture surfaces by generating the frictional force through voltage modulation when a finger is sliding on the display surface. However, it is challenging to prepare and fine-tune the appropriate frictional signals for haptic design and texture simulation. We present FrictGAN, a deep-learningbased framework to synthesize frictional signals for electrostatic tactile displays from fabric texture images. Leveraging GANs (Generative Adversarial Networks), FrictGAN could generate the displacement-series data of frictional coefficients for the electrostatic tactile display to simulate the tactile feedback of fabric material. Our preliminary experimental results showed that FrictGAN could achieve considerable performance on frictional signal generation based on the input images of fabric textures.

Original language	English
Title of host publication	ICAT-EGVE 2020
Subtitle of host publication	International Conference on Artificial Reality and Telexistence and Eurographics Symposium on Virtual Environments
Publisher	The Eurographics Association
Pages	11-15
Number of pages	5
ISBN (Print)	978-3-03868-111-3
DOIs	https://doi.org/10.2312/egve.20201254
Publication status	Published - Dec 2020
Event	ICAT-EGVE2020: International Conference on Artificial Reality and Telexistence & Eurographics Symposium on Virtual Environments - Virtual Duration: 2 Dec 2020 → 4 Dec 2020

Conference

Conference	ICAT-EGVE2020
Period	2/12/20 → 4/12/20

Research Keywords

Generative adversarial network (GAN)
Virtual Reality

Access Output

10.2312/egve.20201254

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Haptic Modeling and Rendering Techniques for Material Simulation and Modulation in Virtual and Mixed Reality
CAI, S. (Author), ZHU, K. (Supervisor), 14 Aug 2023
Student thesis: Doctoral Thesis

Cite this

@inproceedings{b8b4f82414e0426b953c672719acb1cf,

title = "FrictGAN: Frictional Signal Generation from Fabric Texture Images using Generative Adversarial Network",

abstract = "The electrostatic tactile display could render the tactile feeling of different haptic texture surfaces by generating the frictional force through voltage modulation when a finger is sliding on the display surface. However, it is challenging to prepare and fine-tune the appropriate frictional signals for haptic design and texture simulation. We present FrictGAN, a deep-learningbased framework to synthesize frictional signals for electrostatic tactile displays from fabric texture images. Leveraging GANs (Generative Adversarial Networks), FrictGAN could generate the displacement-series data of frictional coefficients for the electrostatic tactile display to simulate the tactile feedback of fabric material. Our preliminary experimental results showed that FrictGAN could achieve considerable performance on frictional signal generation based on the input images of fabric textures.",

keywords = "Generative adversarial network (GAN), Virtual Reality",

author = "Shaoyu Cai and Yuki Ban and Takuji Narumi and Kening Zhu",

year = "2020",

month = dec,

doi = "10.2312/egve.20201254",

language = "English",

isbn = "978-3-03868-111-3",

pages = "11--15",

booktitle = "ICAT-EGVE 2020",

publisher = "The Eurographics Association",

note = "ICAT-EGVE2020 : International Conference on Artificial Reality and Telexistence & Eurographics Symposium on Virtual Environments ; Conference date: 02-12-2020 Through 04-12-2020",

}

Cai, S, Ban, Y, Narumi, T & Zhu, K 2020, FrictGAN: Frictional Signal Generation from Fabric Texture Images using Generative Adversarial Network. in ICAT-EGVE 2020: International Conference on Artificial Reality and Telexistence and Eurographics Symposium on Virtual Environments. The Eurographics Association, pp. 11-15, ICAT-EGVE2020, 2/12/20. https://doi.org/10.2312/egve.20201254

FrictGAN: Frictional Signal Generation from Fabric Texture Images using Generative Adversarial Network. / Cai, Shaoyu; Ban, Yuki; Narumi, Takuji et al.
ICAT-EGVE 2020: International Conference on Artificial Reality and Telexistence and Eurographics Symposium on Virtual Environments. The Eurographics Association, 2020. p. 11-15.

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - FrictGAN

T2 - ICAT-EGVE2020

AU - Cai, Shaoyu

AU - Ban, Yuki

AU - Narumi, Takuji

AU - Zhu, Kening

PY - 2020/12

Y1 - 2020/12

N2 - The electrostatic tactile display could render the tactile feeling of different haptic texture surfaces by generating the frictional force through voltage modulation when a finger is sliding on the display surface. However, it is challenging to prepare and fine-tune the appropriate frictional signals for haptic design and texture simulation. We present FrictGAN, a deep-learningbased framework to synthesize frictional signals for electrostatic tactile displays from fabric texture images. Leveraging GANs (Generative Adversarial Networks), FrictGAN could generate the displacement-series data of frictional coefficients for the electrostatic tactile display to simulate the tactile feedback of fabric material. Our preliminary experimental results showed that FrictGAN could achieve considerable performance on frictional signal generation based on the input images of fabric textures.

AB - The electrostatic tactile display could render the tactile feeling of different haptic texture surfaces by generating the frictional force through voltage modulation when a finger is sliding on the display surface. However, it is challenging to prepare and fine-tune the appropriate frictional signals for haptic design and texture simulation. We present FrictGAN, a deep-learningbased framework to synthesize frictional signals for electrostatic tactile displays from fabric texture images. Leveraging GANs (Generative Adversarial Networks), FrictGAN could generate the displacement-series data of frictional coefficients for the electrostatic tactile display to simulate the tactile feedback of fabric material. Our preliminary experimental results showed that FrictGAN could achieve considerable performance on frictional signal generation based on the input images of fabric textures.

KW - Generative adversarial network (GAN)

KW - Virtual Reality

U2 - 10.2312/egve.20201254

DO - 10.2312/egve.20201254

M3 - RGC 32 - Refereed conference paper (with host publication)

SN - 978-3-03868-111-3

SP - 11

EP - 15

BT - ICAT-EGVE 2020

PB - The Eurographics Association

Y2 - 2 December 2020 through 4 December 2020

ER -

FrictGAN: Frictional Signal Generation from Fabric Texture Images using Generative Adversarial Network

Abstract

Conference

Research Keywords

Access Output

Other Access Options

Permanent Link

Fingerprint

Student theses

Haptic Modeling and Rendering Techniques for Material Simulation and Modulation in Virtual and Mixed Reality

Cite this