PtychoEP: Expectation Propagation for Ptychography
Reproducible implementation of the numerical experiments from our paper A Message-Passing Perspective on Ptychographic Phase Retrieval (IEEE TCI, 2025).
Software
gPIE: Graph-based Probabilistic Inference Engine
gPIE is a modular, extensible Python framework for Expectation Propagation (EP) on factor graphs, tailored for inference problems in computational imaging. It provides a DSL syntax to describe probabilistic models, and seamless NumPy/CuPy backend switching for GPU acceleration.