@inproceedings{e6a41623d16b406181874a5f875172f7,
title = "Generative Adversarial Networks method of polarization 3D reconstruction based on meta-transfer learning",
abstract = "This paper makes a first attempt to combines meta-transfer learning and generative adversarial networks (GAN) technology to solve the Shape from Polarization (SfP) problem. To solve this physics-based ill-posed problem, some researchers choose to blend these physical models as priors into a neural network architecture however cannot meet the requirements of cross-domain and few-shot polarization data. This proposed approach put forward two innovative points. First, we design the meta-transfer method so adapt GAN for few-shot learning tasks. Second, we introduce physical priors between monocular polarization sequence and 3D normal vector into the generative loss term. We report to exceed the previous state-of-the-art on deepsfp dataset, showing the potential of meta-transfer learning in few-shot generative tasks.",
keywords = "Shape from Polarization, generative adversarial networks, meta-transfer learning",
author = "Yuxuan Mao and Kun Gao",
note = "Publisher Copyright: {\textcopyright} 2023 SPIE.; 9th Symposium on Novel Photoelectronic Detection Technology and Applications ; Conference date: 21-04-2023 Through 23-04-2023",
year = "2023",
doi = "10.1117/12.2663312",
language = "English",
series = "Proceedings of SPIE - The International Society for Optical Engineering",
publisher = "SPIE",
editor = "Junhao Chu and Wenqing Liu and Hongxing Xu",
booktitle = "Ninth Symposium on Novel Photoelectronic Detection Technology and Applications",
address = "United States",
}