Finding the semantic similarity in single-particle diffraction images using self-supervised contrastive projection learning

作     者：Julian Zimmermann Fabien Beguet Daniel Guthruf Bruno Langbehn Daniela Rupp 

作者机构：ETH ZürichZürichSwitzerland Technische Universität BerlinBerlinGermany Max-Born-InstitutBerlinGermany 

出 版 物：《npj Computational Materials》 (计算材料学（英文）)

年 卷 期：2023年第9卷第1期

页      面：2118-2126页

核心收录：

学科分类：07[理学] 070205[理学-凝聚态物理] 0804[工学-仪器科学与技术] 0702[理学-物理学] 

基　　金：Leibniz Society, (SAW/2017/MBI4) NCCR MUST Deutsche Forschungsgemeinschaft, DFG, (MO 719/14-2) Deutsche Forschungsgemeinschaft, DFG Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung, SNF, (200021E_193642) Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung, SNF Eidgenössische Technische Hochschule Zürich, ETH 

主　　题：projection coherent similarity 

摘      要：Single-shot coherent diffraction imaging of isolated nanosized particles has seen remarkable success in recent years,yielding in-situ measurements with ultra-high spatial and temporal resolution.The progress of high-repetition-rate sources for intense X-ray pulses has further enabled recording datasets containing millions of diffraction images,which are needed for the structure determination of specimens with greater structural variety and dynamic experiments.The size of the datasets,however,represents a monumental problem for their analysis.Here,we present an automatized approach for finding semantic similarities in coherent diffraction images without relying on human expert labeling.By introducing the concept of projection learning,we extend self-supervised contrastive learning to the context of coherent diffraction imaging and achieve a dimensionality reduction producing semantically meaningful embeddings that align with physical intuition.The method yields substantial improvements compared to previous approaches,paving the way toward real-time and large-scale analysis of coherent diffraction experiments at X-ray free-electron lasers.