PET-Adapter: Test-Time Domain Adaptation for Full and Limited-Angle PET Image Reconstruction
2026-05-08 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionMachine Learning
AI summaryⓘ
The authors address challenges in PET imaging caused by noisy data and limited viewing angles, which make it hard to get clear pictures. They created PET-Adapter, a method that adapts pretrained models (trained on simple phantom images) to work well with real clinical data without needing matched examples. Their technique uses information about body anatomy and starts with physics-based reconstructions to make the process faster and still accurate. Tests showed that their approach improves 3D image quality and works efficiently across different clinical setups.
Positron Emission Tomography (PET)image reconstructionPoisson noiselimited-angle acquisitiondomain adaptationphantom dataanatomical conditioningOrdered Subset Expectation Maximization (OSEM)diffusion models3D reconstruction
Authors
Rüveyda Yilmaz, Yuli Wu, Johannes Stegmaier, Volkmar Schulz
Abstract
Positron Emission Tomography (PET) image reconstruction is inherently challenged by Poisson noise and physical degradation factors, which are further exacerbated in limited-angle acquisitions. While deep learning methods demonstrate promising performance, their generalization to unseen clinical data distributions remains limited without extensive retraining. We propose PET-Adapter, a test-time domain adaptation framework for generative PET reconstruction models pretrained solely on phantom data. Our method enables adaptation to clinical datasets with varying anatomies, tracers, and scanner configurations without requiring paired ground truth. PET-Adapter introduces layer-wise low-rank anatomical conditioning during adaptation and Ordered Subset Expectation Maximization-based warm-starting that initializes the generation from physics-informed reconstructions, reducing diffusion steps from 50 to 2 without compromising quality. Experiments across multiple clinical datasets demonstrate superior 3D reconstruction performance in both full-angle and limited-angle settings, highlighting the clinical feasibility and computational efficiency of the proposed approach.