Three-dimensional coherent X-ray diffraction imaging via deep convolutional neural networks

Abstract As a critical component of coherent X-ray diffraction imaging (CDI), phase retrieval has been extensively applied in structural science to recover the 3D morphological information inside measured particles. Despite meeting all oversampling requirements Sayre and Shannon, current approaches still have trouble achieving unique inversion experimental data presence noise. Here, we propose overcome this limitation by incorporating Machine Learning (ML) model combining (optional) supervised learning with transfer learning. The trained ML can rapidly provide an immediate result high accuracy which could benefit real-time experiments, predicted be further refined More significantly, proposed used without any prior training learn missing phases image based on minimization appropriate ‘loss function’ alone. We demonstrate significantly improved performance Bragg CDI over traditional iterative algorithms.