TiAM-GAN: Titanium Alloy Microstructure Image Generation Network

Zhixuan Zhang, Fusheng Jin*, Haichao Gong, Qunbo Fan

*此作品的通讯作者

科研成果: 书/报告/会议事项章节会议稿件同行评审

摘要

The generation of titanium alloy microstructure images through mechanical properties is of great value to the research and production of titanium alloy materials. The appearance of GAN provides the possibility for image generation. However, there is currently no work related to handling multiple continuous labels for microstructure images. This paper presents a multi-label titanium alloy microstructure image generation network(TiAM-GAN). The TiAM-GAN proposed in this paper contains two sub-networks, a generation network for simple textures, which is based on the existing generation adversarial network, we reconstruct the loss function for multi-label continuous variables and deduce the error bound. Another microstructure image generation network for complex textures uses a mixture density network to learn the labels-to-noise mapping, and a deep convolution generation adversarial network is used to learn the noise-to-image mapping, then the noise output by the mixture density network will input to the deep convolution generation adversarial network to generate the image. Finally, we compared our method with existing methods qualitatively and quantitatively, which shows our method can achieve better results.

源语言英语
主期刊名Pattern Recognition and Computer Vision - 6th Chinese Conference, PRCV 2023, Proceedings
编辑Qingshan Liu, Hanzi Wang, Rongrong Ji, Zhanyu Ma, Weishi Zheng, Hongbin Zha, Xilin Chen, Liang Wang
出版商Springer Science and Business Media Deutschland GmbH
84-96
页数13
ISBN(印刷版)9789819984343
DOI
出版状态已出版 - 2024
活动6th Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2023 - Xiamen, 中国
期限: 13 10月 202315 10月 2023

出版系列

姓名Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
14427 LNCS
ISSN(印刷版)0302-9743
ISSN(电子版)1611-3349

会议

会议6th Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2023
国家/地区中国
Xiamen
时期13/10/2315/10/23

指纹

探究 'TiAM-GAN: Titanium Alloy Microstructure Image Generation Network' 的科研主题。它们共同构成独一无二的指纹。

引用此