Share Email Print
cover

Proceedings Paper

A simulation model based on DCGAN to generate 2D animation avatars
Author(s): Keting Chen; Ce Gao; Yiran Cai
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

The rising of machine learning has enabled computers to do difficult tasks that once and only be don e by human beings. However, many applications haven’t taught the machine about the field of art. To make machine more artistic, this paper designed a model based on GAN that enables the computer to generate 2D Japanese animation avatar. In the paper, GAN is found to be an effective method: the generator and discriminator work together to generate normal but real pictures. The experimental result is that in the 300 generated animations, only 3 images are distorted or unreal, indicating that the success rate is 99% and the proposed model has achieved excellent performance. Furthermore, it can be found that most of the generated pictures are beautiful, meaning that the machine is able to draw not only real, but also beautiful pictures.

Paper Details

Date Published: 28 July 2022
PDF: 7 pages
Proc. SPIE 12303, International Conference on Cloud Computing, Internet of Things, and Computer Applications (CICA 2022), 1230324 (28 July 2022); doi: 10.1117/12.2642603
Show Author Affiliations
Keting Chen, University of Wisconsin-Madison (United States)
Ce Gao, Jiangnan University (China)
Yiran Cai, University of Sydney (Australia)


Published in SPIE Proceedings Vol. 12303:
International Conference on Cloud Computing, Internet of Things, and Computer Applications (CICA 2022)
Warwick Powell; Amr Tolba, Editor(s)

© SPIE. Terms of Use
Back to Top
PREMIUM CONTENT
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?
close_icon_gray