PtyGenography: using generative models for regularization of the phase retrieval problem
This work addresses stability issues in inverse problems for applications like imaging, but it appears incremental as it builds on existing generative regularization methods.
The paper tackled the instability of solutions in phase retrieval and similar inverse problems by proposing a unified reconstruction approach that mitigates overfitting to generative models used as regularization, comparing it with classical formulations.
In phase retrieval and similar inverse problems, the stability of solutions across different noise levels is crucial for applications. One approach to promote it is using signal priors in a form of a generative model as a regularization, at the expense of introducing a bias in the reconstruction. In this paper, we explore and compare the reconstruction properties of classical and generative inverse problem formulations. We propose a new unified reconstruction approach that mitigates overfitting to the generative model for varying noise levels.