Less Is More: Training-Free Acceleration Framework of 3D Diffusion Models for Low-Count PET Denoising via Global-Local Trajectory Reduction
About
Accurate quantification and uptake measurement in PET are critical for assessing disease progression and supporting clinical decision-making. While high-count PET provides reliable image quality, the associated radiation dose and prolonged acquisition remain significant clinical concerns, motivating the adoption of low-count protocols. Diffusion-model-based methods have demonstrated strong potential for restoring low-count PET to near high-count quality, but their iterative sampling procedure becomes prohibitively expensive when applied to high-resolution 3D PET volumes, introducing substantial inference latency that limits practical clinical deployment. To address these challenges, we propose a training-free Global-Local Skipping Strategy that accelerates diffusion model-based 3D PET denoising while simultaneously improving reconstruction quality. The proposed method is plug-and-play and directly applicable to pre-trained diffusion models without retraining or architectural modification. Specifically, we introduce: (i) a global denoising step skipping strategy that initializes the reverse diffusion process from an intermediate denoising step using a noise-consistent transformation of the low-count input, substantially reducing the number of required denoising steps; and (ii) a local feature reuse shortcut that reuses slowly-varying high-level U-Net features across neighboring denoising steps, further reducing per-step computation while preserving image fidelity. We evaluate the proposed approach on multiple PET tracers from in-house and public datasets, including 18F-FDG PET, 68Ga-DOTATATE PET, and 18F-PSMA PET, demonstrating consistent acceleration of over an order of magnitude alongside improved or comparable reconstruction performance relative to the full-step baseline. Blinded reader studies further confirm enhanced clinical confidence and perceived diagnostic quality.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| 3D PET Image Denoising | UDPET 18F-FDG PET public (test) | PSNR41.05 | 7 | |
| PET Image Restoration | 18F-FDG 15 Cases | PSNR42.18 | 7 | |
| PET Image Restoration | 68Ga-DOTATATE (32 Cases) | PSNR43.85 | 7 | |
| PET Image Restoration | 18F-PSMA (25 Cases) | PSNR44.78 | 7 | |
| PET Denoising | 18F-FDG PET cohort (Northwestern Memorial Hospital) (in-house) | Inference Time (min)1.3 | 6 | |
| PET Denoising | 68Ga-DOTATATE PET cohort (Northwestern Memorial Hospital) (in-house) | Inference Time (min)1 | 6 | |
| PET Denoising | 18F-PSMA PET cohort Northwestern Memorial Hospital (in-house) | Inference Time (min)0.3 | 6 |