Research Article
A Deep Learning-Based Approach for Generating 3D Models of Fluid Arts
@INPROCEEDINGS{10.1007/978-3-031-28993-4_4, author={Hung Mai Cong and Mai Xuan Trang and Akihiro Yamada and Suzuki Takashi and Naoko Tosa and Ryohei Nakatsu}, title={A Deep Learning-Based Approach for Generating 3D Models of Fluid Arts}, proceedings={ArtsIT, Interactivity and Game Creation. 11th EAI International Conference, ArtsIT 2022, Faro, Portugal, November 21-22, 2022, Proceedings}, proceedings_a={ARTSIT}, year={2023}, month={4}, keywords={3D fluid art Sound of Ikebana DIB-R network GAN variants}, doi={10.1007/978-3-031-28993-4_4} }- Hung Mai Cong
Mai Xuan Trang
Akihiro Yamada
Suzuki Takashi
Naoko Tosa
Ryohei Nakatsu
Year: 2023
A Deep Learning-Based Approach for Generating 3D Models of Fluid Arts
ARTSIT
Springer
DOI: 10.1007/978-3-031-28993-4_4
Abstract
This paper explores a method for creating 3D models of video artwork based on a fluid phenomenon called the Sound of Ikebana artwork. A process of using multiple generative adversarial networks (GANs) to reconstruct and predict the shape of the fluid artworks from two-dimensional reference photos was proposed. This is an extension of our previous efforts with Wasserstein GAN enhancements to predict the shape of the unmapped part and correct the texture. The experiment’s results show that our process can reconstruct 3D arts without having large amount of 3D training data.
Copyright © 2022–2025 ICST
