AutoPhaseNN: Unsupervised Physics-aware Deep Learning of 3D Nanoscale Coherent Imaging

The problem of phase retrieval, or the algorithmic recovery of lost phase information from measured intensity alone, underlies various imaging methods from astronomy to nanoscale imaging. Traditional methods of phase retrieval are iterative in nature, and are therefore computationally expensive and timeconsuming. More recently, deep learning (DL) models have been developed to either provide learned priors to iterative phase retrieval or in some cases completely replace phase retrieval with networks that learn to recover the lost phase information from measured intensity alone. However, such models require vast amounts of labeled data, which can only be obtained through simulation or performing computationally prohibitive phase retrieval on hundreds of or even thousands of experimental datasets. Using a 3D nanoscale X-ray imaging modality (Bragg Coherent Diffraction Imaging or BCDI) as a representative technique, we demonstrate AutoPhaseNN, a DL-based approach

[1]  Y. S. Meng,et al.  Topological defect dynamics in operando battery nanoparticles , 2015, Science.

[2]  J. Miao,et al.  Beyond crystallography: Diffractive imaging using coherent x-ray light sources , 2015, Science.

[3]  Youssef S. G. Nashed,et al.  Real-time coherent diffraction inversion using deep generative networks , 2018, Scientific Reports.

[4]  Ian McNulty,et al.  Ultrafast Three-Dimensional X-ray Imaging of Deformation Modes in ZnO Nanocrystals. , 2017, Nano letters.

[5]  Oren Cohen,et al.  Deep neural networks in single-shot ptychography. , 2020, Optics express.

[6]  D. Scott Acton,et al.  Phase retrieval algorithm for JWST Flight and Testbed Telescope , 2006, SPIE Astronomical Telescopes + Instrumentation.

[7]  R Harder,et al.  High-resolution three-dimensional partially coherent diffraction imaging , 2012, Nature Communications.

[8]  Ian K. Robinson,et al.  Three-dimensional imaging of dislocation propagation during crystal growth and dissolution , 2015, Nature materials.

[9]  Justin S. Wark,et al.  Ultrafast Three-Dimensional Imaging of Lattice Dynamics in Individual Gold Nanocrystals , 2013, Science.

[10]  Kyoung Mu Lee,et al.  Enhanced Deep Residual Networks for Single Image Super-Resolution , 2017, 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[11]  R. Pokharel,et al.  Adaptive 3D convolutional neural network-based reconstruction method for 3D coherent diffraction imaging , 2020, 2008.10094.

[12]  Ian McNulty,et al.  Ultrafast Three-Dimensional Integrated Imaging of Strain in Core/Shell Semiconductor/Metal Nanostructures. , 2017, Nano letters.

[13]  D. Ratner,et al.  Recovering the phase and amplitude of X-ray FEL pulses using neural networks and differentiable models. , 2021, Optics express.

[14]  Youssef S. G. Nashed,et al.  AI-enabled high-resolution scanning coherent diffraction imaging , 2020 .

[15]  Stefan Harmeling,et al.  Image denoising: Can plain neural networks compete with BM3D? , 2012, 2012 IEEE Conference on Computer Vision and Pattern Recognition.

[16]  M. Graef,et al.  Recent advances in Lorentz microscopy , 2016 .

[17]  Wonsuk Cha,et al.  Three-dimensional X-ray diffraction imaging of dislocations in polycrystalline metals under tensile loading , 2018, Nature Communications.

[18]  Veit Elser,et al.  Electron ptychography of 2D materials to deep sub-ångström resolution , 2018, Nature.

[19]  Ian K. Robinson,et al.  3D lattice distortions and defect structures in ion-implanted nano-crystals , 2017, Scientific Reports.

[20]  Christian Ledig,et al.  Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[21]  Kamel Fezzaa,et al.  PhaseGAN: a deep-learning phase-retrieval approach for unpaired datasets. , 2020, Optics express.

[22]  Xiaojing Huang,et al.  PtychoNet: Fast and High Quality Phase Retrieval for Ptychography , 2019, BMVC.

[23]  Abhay Patil,et al.  Learning Image Restoration without Clean Data , 2019 .

[24]  J. Miao,et al.  The oversampling phasing method. , 2000, Acta crystallographica. Section D, Biological crystallography.

[25]  P. Thibault.  X-ray ptychography , 2011 .

[26]  R. Harder,et al.  Real-time 3D Nanoscale Coherent Imaging via Physics-aware Deep Learning , 2020, ArXiv.

[27]  Hanfei Yan,et al.  Dynamic diffraction artefacts in Bragg coherent diffractive imaging , 2018, Journal of applied crystallography.

[28]  Ian K Robinson,et al.  Imaging transient melting of a nanocrystal using an X-ray laser , 2015, Proceedings of the National Academy of Sciences.

[29]  P Zapol,et al.  Avalanching strain dynamics during the hydriding phase transformation in individual palladium nanoparticles , 2015, Nature Communications.

[30]  Lei Tian,et al.  Deep learning approach for Fourier ptychography microscopy. , 2018, Optics express.

[31]  Ian K. Robinson,et al.  Complex imaging of phase domains by deep neural networks , 2021, IUCrJ.

[32]  Xiaoou Tang,et al.  Image Super-Resolution Using Deep Convolutional Networks , 2014, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[33]  Yibo Zhang,et al.  Phase recovery and holographic image reconstruction using deep learning in neural networks , 2017, Light: Science & Applications.

[34]  R. Harder.  Deep neural networks in real-time coherent diffraction imaging , 2021, IUCrJ.

[35]  Jesse N. Clark,et al.  Coherent diffraction imaging of nanoscale strain evolution in a single crystal under high pressure , 2013, Nature Communications.

[36]  Elias Vlieg,et al.  Angle calculations for a six‐circle surface X‐ray diffractometer , 1993 .

[37]  Shuai Li,et al.  Lensless computational imaging through deep learning , 2017, ArXiv.