# Contrast-invariant deep ptychography neural networks

arXiv: https://arxiv.org/abs/2608.02869 · HTML: https://mlsanity.com/paper/2608.02869/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 27 references that are in the map.
- First posted on arXiv 3 Aug 2026.

## Builds on (most cited first, 8 of 8)

- [Learning to See by Looking at Noise](https://mlsanity.com/paper/2106.05963/) (2021)
- [On the use of deep learning for phase recovery](https://mlsanity.com/paper/2308.00942/) (2023)
- [Physics Constrained Unsupervised Deep Learning for Rapid, High Resolution Scanning Coherent Diffraction Reconstruction](https://mlsanity.com/paper/2306.11014/) (2023)
- [PtychoDV: Vision Transformer-Based Deep Unrolling Network for Ptychographic Image Reconstruction](https://mlsanity.com/paper/2310.07504/) (2023)
- [PtychoFormer: A Transformer-based Model for Ptychographic Phase Retrieval](https://mlsanity.com/paper/2410.17377/) (2024)
- [Deep-learning real-time phase retrieval of imperfect diffraction patterns from X-ray free-electron lasers](https://mlsanity.com/paper/2409.15784/) (2024)
- [Ptychographic Image Reconstruction from Limited Data via Score-Based Diffusion Models with Physics-Guidance](https://mlsanity.com/paper/2502.18767/) (2025)
- [Towards generalizable deep ptychography neural networks](https://mlsanity.com/paper/2509.25104/) (2025)

## Abstract (arXiv)

Ptychography neural networks suffer from scaling inconsistencies when generalizing out of distribution, limiting their real world viability. We address this scaling mismatch using a factorization strategy which decouples the learned object texture from measurement scaling, enabling a single trained network to produce measurement-consistent reconstructions across varying illumination conditions. This requires predicting the learned object in real and imaginary units instead of the canonical amplitude and phase representation. We additionally introduce a synthetic object sampling strategy that minimizes phase distribution mismatch between synthetic training data and experimental targets. These improvements yield up to a 5x reduction in Fourier error over the previous PtychoPINN-torch baseline across 5 experimental datasets spanning multiple beamlines and facilities.