Abstract
Conventional deep learning-based image reconstruction methods require a large amount of training data, which can be hard to obtain in practice. Untrained deep learning methods overcome this limitation by training a network to invert a physical model of the image formation process. Here we present a novel, to our knowledge, untrained Res-U2Net model for phase retrieval. We use the extracted phase information to determine changes in an object’s surface and generate a mesh representation of its 3D structure. We compare the performance of Res-U2Net phase retrieval against UNet and U2Net using images from the GDXRAY dataset.
© 2024 Optica Publishing Group
Full Article | PDF ArticleMore Like This
Zehua Wang, Shenghao Zheng, Zhihui Ding, and Cheng Guo
J. Opt. Soc. Am. A 41(2) 165-173 (2024)
Kaijun Shi, Xing Zhang, Xin Wang, Jie Xu, Baozhong Mu, Ji Yan, Feng Wang, Yongkun Ding, and Zhanshan Wang
Opt. Express 32(8) 14356-14376 (2024)
Yongfu Zhao, Ruiming Kong, Fei Ma, Sumin Qi, Cuixia Dai, and Jing Meng
Opt. Express 32(10) 17318-17335 (2024)