AI & ML
Stochastic Optimisation for Scalable Ptychographic Image Reconstruction
arXiv eess.IV0 views
arXiv:2611.24232v1 Announce Type: new Abstract: X-ray ptychography is a lensless coherent diffraction imaging technique capable of high spatial resolution, in which an image is reconstructed by solving a large-scale inverse problem from highly redundant far-field diffraction measurements acquired under overlapping illumination. Whilst the measurement redundancy created by overlapping illumination is essential for reconstruction fidelity, it leads to computational and memory demands that scale poorly with object size and target resolution when using conventional reconstruction algorithms. This work proposes a ptychographic reconstruction framework that reduces peak memory usage and time to convergence by leveraging stochastic optimisation techniques developed for large-scale machine learning. Specifically, the framework employs stochastic gradient descent with a decaying step size that progressively reduces the effect of gradient noise, referred to here as variance reduction (VR), with gradients that agree with automatic differentiation, to optimise the ptychographic forward model using randomly sampled subsets of diffraction measurements. For an experimental dataset of 200,000 diffraction patterns of $128\times128$ pixels, VR is demonstrated to require $65\%$ of the peak GPU memory and $22\%$ of the reconstruction time of the established maximum-likelihood (ML) algorithm. The VR framework incorporates the Poisson maximum-likelihood formulation and partial coherence within a unified and extensible optimisation setting. Numerical experiments demonstrate that the framework achieves reconstruction quality visually comparable to established methods, with lower peak GPU memory and a shorter time
Source:https://arxiv.org/abs/2611.24232PreviousMicrograph B-factor estimation via Wilson statisticsNextTransferable Named Sparse Features for Morphology-based Gleason Grading
Related posts
"You’re right"—LLMs will get climate info wrong if you ask
0Ars Technica

Oh, Elon Musk is ‘moderate’ now, is he?
0The Verge
App Store & RevenueCat Fees: 15% vs 30% Tiers (The True Net Margin Guide)
0dev.to
Are CLAUDE.md and AGENTS.md a revival of the design document
0dev.to
[ Video ] Local AI: Testing Domyn Edge 4B on a 16GB MacBook Pro
0dev.to

Enabling Real-Time AI for Scientific Instruments with GPU-Native Data Acquisition
0NVIDIA Developer Blog