Score-Based Generative Models for PET Image Reconstruction

Imraj RD Singh*,10000-0003-2186-0977, Alexander Denker*,20000-0002-7265-261X, Riccardo Barbano*,10000-0003-1863-2092, Željko Kereta10000-0003-2805-0037, Bangti Jin30000-0002-3775-9155, Kris Thielemans40000-0002-5514-199X, Peter Maass20000-0003-1448-8345, Simon Arridge10000-0003-1292-0210
*: Equal contribution, 1: Department of Computer Science, University College London, 2: Center for Industrial Mathematics, University of Bremen, 3: Department of Mathematics, Chinese University of Hong Kong, 4: Institute of Nuclear Medicine, University College London
Publication date: 2024/01/23
PDF · Code · Dataset · arXiv


Score-based generative models have demonstrated highly promising results for medical image reconstruction tasks in magnetic resonance imaging or computed tomography. However, their application to Positron Emission Tomography (PET) is still largely unexplored. PET image reconstruction involves a variety of challenges, including Poisson noise with high variance and a wide dynamic range. To address these challenges, we propose several PET-specific adaptations of score-based generative models. The proposed framework is developed for both 2D and 3D PET. In addition, we provide an extension to guided reconstruction using magnetic resonance images. We validate the approach through extensive 2D and 3D in-silico experiments with a model trained on patient-realistic data without lesions, and evaluate on data without lesions as well as out-of-distribution data with lesions. This demonstrates the proposed method’s robustness and significant potential for improved PET reconstruction.


Positron Emission Tomography · Diffusion models · Score-based generative models · Image Reconstruction · 3D image reconstruction · Guided reconstruction

Bibtex @article{melba:2024:001:singh, title = "Score-Based Generative Models for PET Image Reconstruction", author = "Singh, Imraj RD and Denker, Alexander and Barbano, Riccardo and Kereta, Željko and Jin, Bangti and Thielemans, Kris and Maass, Peter and Arridge, Simon", journal = "Machine Learning for Biomedical Imaging", volume = "2", issue = "Special Issue for Generative Models", year = "2024", pages = "547--585", issn = "2766-905X", doi = "", url = "" }
RISTY - JOUR AU - Singh, Imraj RD AU - Denker, Alexander AU - Barbano, Riccardo AU - Kereta, Željko AU - Jin, Bangti AU - Thielemans, Kris AU - Maass, Peter AU - Arridge, Simon PY - 2024 TI - Score-Based Generative Models for PET Image Reconstruction T2 - Machine Learning for Biomedical Imaging VL - 2 IS - Special Issue for Generative Models SP - 547 EP - 585 SN - 2766-905X DO - UR - ER -

2024:001 cover