External phantom-based validation of a deep-learning network trained for upscaling of digital low count PET data
Phantom-based validation of AI-based algorithms allows for a detailed assessment of the performance, limitations, and generalizability of deep-learning based algorithms for PET image enhancement. Here it was confirmed that the AI-based approach performs very well in denoising ultra-low count PET images and outperforms traditional Gaussian post-filtering. However, there are strong limitations in terms of quantitative accuracy and detectability of small lesions.
Authors
- Robert Seifert (ORCID: https://orcid.org/0000-0001-5985-7701)
- Anja Braune (ORCID: https://orcid.org/0000-0001-7707-9413)
- René Hosch (ORCID: https://orcid.org/0000-0003-1760-2342)
- David Kersting (ORCID: https://orcid.org/0000-0002-8451-1830)
- Juliane Müller (ORCID: https://orcid.org/0000-0002-4966-8285)
- Frank Hofheinz (ORCID: https://orcid.org/0000-0001-8016-4643)
- Felix Nensa (ORCID: https://orcid.org/0000-0002-5811-7100)
- Ken Herrmann
- Jörg Kotzerke
Institutions
- University of Bern (CH)
- University Hospital of Bern (CH)
- Helmholtz-Zentrum Dresden-Rossendorf (DE)
- Essen University Hospital (DE)
- University Hospital Carl Gustav Carus (DE)
Publication Details
- Journal
- EJNMMI Physics
- Published
- 2025-04-16
- DOI
- https://doi.org/10.1186/s40658-025-00745-4
- Citations
- 2
- Primary Topic
- Medical Imaging Techniques and Applications
- Type
- article
- Field-Weighted Citation Impact
- 2.72