ConvexAD: differentiable phase retrieval with geometric regularization for high-strain, low-signal Bragg coherent diffraction imaging
Abstract
Bragg Coherent Diffraction Imaging (BCDI) relies on phase retrieval (PR), a non-convex, ill-posed inverse problem whose conditioning is degraded by factors like noise, partial coherence, and strain or crystallographic defects in the imaged nanocrystal. Conventional approaches, including alternating projections and deep learning methods, represent the object on a dense voxel grid, which is prone to amplitude-phase mixing and becomes increasingly ill-posed under these factors, often preventing reliable three-dimensional mapping of the particle's shape and strain. Here we present ConvexAD, a gradient descent PR model that parametrizes the object shape as a convex polyhedron, built by the intersection of half-spaces and enforces flat amplitude within it by construction. This geometric regularization eliminates amplitude-phase mixing and reduces the number of trainable parameters, hence facilitating the convergence of the PR. Implemented in JAX for native GPU acceleration and forward-model flexibility, ConvexAD is validated on simulated and experimental BCDI data, where it outperforms conventional and deep learning assisted PR at high strain, low signal, and reduced coherence, extending the conditions under which quantitative BCDI reconstruction is achievable.
Related
- Differentiable powder X-ray diffraction enables analysis of evolving structures
- Image Priors for Single-Shot Coded Diffraction Phase Retrieval
- Beyond the Thin-Layer Limit: Differentiable Volumetric Training for Visible-Range Diffractive Neural Networks
- Generalizable data-driven turbulence closure modeling on unstructured grids with differentiable physics
- Efficient angle-aware differentiable Mask3D modeling for inverse lithography
BibTeX
@misc{masto2026convexad,
title = {{ConvexAD: differentiable phase retrieval with geometric regularization for high-strain, low-signal Bragg coherent diffraction imaging}},
author = {Matteo Masto and Ewen Bellec and Etienne Polack and Jiangtao Zhao and Vincent Favre-Nicolin and Steven J. Leake and Tobias Schülli and Marie-Ingrid Richard},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.05020},
eprint = {2611.05020},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.05020},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - ConvexAD: differentiable phase retrieval with geometric regularization for high-strain, low-signal Bragg coherent diffraction imaging
AU - Matteo Masto
AU - Ewen Bellec
AU - Etienne Polack
AU - Jiangtao Zhao
AU - Vincent Favre-Nicolin
AU - Steven J. Leake
AU - Tobias Schülli
AU - Marie-Ingrid Richard
PY - 2026
DA - 2026/11/03
PB - arXiv
DO - 10.48550/arxiv.2611.05020
UR - https://arxiv.org/abs/2611.05020
AB - Bragg Coherent Diffraction Imaging (BCDI) relies on phase retrieval (PR), a non-convex, ill-posed inverse problem whose conditioning is degraded by factors like noise, partial coherence, and strain or crystallographic defects in the imaged nanocrystal. Conventional approaches, including alternating projections and deep learning methods, represent the object on a dense voxel grid, which is prone to amplitude-phase mixing and becomes increasingly ill-posed under these factors, often preventing reliable three-dimensional mapping of the particle's shape and strain. Here we present ConvexAD, a gradient descent PR model that parametrizes the object shape as a convex polyhedron, built by the intersection of half-spaces and enforces flat amplitude within it by construction. This geometric regularization eliminates amplitude-phase mixing and reduces the number of trainable parameters, hence facilitating the convergence of the PR. Implemented in JAX for native GPU acceleration and forward-model flexibility, ConvexAD is validated on simulated and experimental BCDI data, where it outperforms conventional and deep learning assisted PR at high strain, low signal, and reduced coherence, extending the conditions under which quantitative BCDI reconstruction is achievable.
ER -