Autor:	Zhong Zhuang, Ju Sun, Kshitij Tayal, Raunak Manekar, Chieh-Hsin Lai, Vipin Kumar
Rok vydání:	2021
Předmět:	Work (thermodynamics) Theoretical computer science Artificial neural network Machine vision Computer science Homogeneous space Learning methods Inverse problem Phase retrieval Data-driven
Zdroj:	Experts@Minnesota
DOI:	10.1364/cosi.2021.cth4a.4
Popis:	The existing iterative and data-driven methods fail to solve phase retrieval due to the intrinsic problem symmetries. We propose two end-to-end learning methods that break the barrier and work in a new regime.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_dedup___::9973f6d894b428854c711138faf71eac https://doi.org/10.1364/cosi.2021.cth4a.4 Zobrazit plný text záznamu