View source for Projects/SRF: Syncrotron radiation facility deep image retrieval

{{#seo:
 |title=Syncrotron radiation facility deep image retrieval
 |titlemode=replace
 |keywords=Syncrotron radiation facility deep image retrieval
 |description=Projects/SRF: Syncrotron radiation facility deep image retrieval
 }}
==Project 58 ==
* Title: X-ray nanotomographyimage reconstruction
* Problem: To boost the quality of nano-object images, obtained in the European Syncrotron Radiation Facility, [https://www.osug.fr/IMG/pdf/zontone_osug_workshop2017id10.pdf ESRF].
* Data: on your request, 3GB
* References:
** [https://arxiv.org/pdf/1809.04626.pdf] Iterative phase retrieval in coherent diffractive imaging: practical issues
** [https://www.nature.com/articles/s41467-019-08635-x#Sec15] X-ray nanotomography of coccolithophores reveals that coccolith mass and segment number correlate with grid size
** [https://www.nature.com/articles/s41598-018-34253-6#Sec14] Lens-free microscopy for 3D + time acquisitions of 3D cell culture
** [https://arxiv.org/pdf/1904.11301.pdf] DEEP ITERATIVE RECONSTRUCTION FOR PHASE RETRIEVAL
** [https://docs.google.com/document/d/1K7bIzU33MSfeUvg3WITRZX0pe3sibbtH62aw42wxsEI/edit?ts=5e42f70e LinkReview]
** AUSPEX is a diagnostic tool for graphical X-ray data analysis, see [https://www.auspex.de/pathol/ common pathologies and their causes]
* Basic:  Gerchberg-Saxton algorithm 
* Method: To boost Gerchberg-Saxton with neural networks. Use the Bayesian approach and set physical models as expert-given prior information
* Novelty: we are developing expert learning method
* Authors: Sergei Grudinin, Yuri Chushkin, and Vadim Strijov