Reconstruction of Ultrashort Pulses Using Deep Neural Networks

Alex Dikopoltsev, Tom Zahavy, Ron Ziv, Ittai Rubinstein, Pavel Sidorenko, Shai Mannor, Oren Cohen, Mordechai Segev

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Ultrashort pulses are used in numerous applications. Characterization of these pulses requires indirect schemes. A popular technique for characterizing ultrashort pulses is frequency-resolved optical gating (FROG) [1]. It is an established and recently mathematically proven [2] approach for full characterization of the amplitude and phase of ultrashort optical pulses. The reconstruction of pulses from these FROG measurements requires an algorithm that can solve a 4th order phase retrieval problem. Such common algorithms, the Principle Component General Projections Algorithm (PCGPA) [3] and recently proposed Ptychographic FROG [4] perform well at high SNR. However, while these algorithms perform well at low SNR after filtering the FROG trace, their performance drops at raw low SNR data.

Original languageEnglish
Title of host publication2018 2nd URSI Atlantic Radio Science Meeting, AT-RASC 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9789082598735
DOIs
StatePublished - 24 Sep 2018
Externally publishedYes
Event2nd URSI Atlantic Radio Science Meeting, AT-RASC 2018 - Gran Canaria, Spain
Duration: 28 May 20181 Jun 2018

Publication series

Name2018 2nd URSI Atlantic Radio Science Meeting, AT-RASC 2018

Conference

Conference2nd URSI Atlantic Radio Science Meeting, AT-RASC 2018
Country/TerritorySpain
CityGran Canaria
Period28/05/181/06/18

Fingerprint

Dive into the research topics of 'Reconstruction of Ultrashort Pulses Using Deep Neural Networks'. Together they form a unique fingerprint.

Cite this