Abstract
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.
| Original language | English |
|---|---|
| Title of host publication | 2025 International Conference on Sampling Theory and Applications (SampTA) |
| Publisher | IEEE |
| DOIs | |
| Publication status | Published - 28 Jul 2025 |
| Event | 2025 International Conference on Sampling Theory and Applications, SampTA 2025 - Vienna, Austria Duration: 28 Jul 2025 → 1 Aug 2025 |
Publication series
| Name | International Conference on Sampling Theory and Applications, SampTA |
|---|---|
| ISSN (Print) | 2831-5480 |
Conference
| Conference | 2025 International Conference on Sampling Theory and Applications, SampTA 2025 |
|---|---|
| Country/Territory | Austria |
| City | Vienna |
| Period | 28/07/25 → 1/08/25 |
Bibliographical note
Publisher Copyright:© 2025 IEEE.
Keywords
- generative priors
- inverse problems
- phase retrieval
- regularization
Fingerprint
Dive into the research topics of 'PtyGenography: Using generative models for regularization of the phase retrieval problem'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver