Research map / Probabilistic and scientific modeling / Molecules and proteins
FlowPET: Physics-Informed Symplectic Flow Matching for Low-Count PET Reconstruction
Zheng Zhang, Hao Tang, Yingying Hu, zhanli hu, Jing Qin
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add restoring range-space momentum p0 (ablation) | FlowPET · UDPET Brain | SSIM: 0.8915 -> 0.9047 | Improved |
| add thermal null-space momentum p1 on top of restoring p0 (ablation) | FlowPET · In-House whole-body | SSIM: 0.9766 -> 0.9811 | Improved |
| add thermal null-space momentum p1 on top of restoring p0 (ablation) | FlowPET · In-House whole-body | PSNR: 36.36 -> 36.35 | No clear effect |
| add thermal null-space momentum p1 on top of restoring p0 (ablation) | FlowPET · UDPET Brain | SSIM: 0.9047 -> 0.9139 | Improved |
| set momentum compression factor gamma to 10^-2 instead of 1 (ablation) | FlowPET · In-House whole-body | SSIM: 0.9616 -> 0.9811 | Improved |
| set gamma to 10^-2 instead of 10^-3 (ablation) | FlowPET · In-House whole-body | SSIM: 0.9796 -> 0.9811 | Improved |
| use FlowPET instead of FourierPET | FlowPET · In-House pediatric (1% count) | SSIM: 0.9740 -> 0.9811 | Improved |
| use FlowPET instead of FourierPET | FlowPET · BrainWeb (20% count) | SSIM: 0.9859 -> 0.9838 | Hurt |
| use FlowPET instead of DREAM | FlowPET · UDPET Brain (1% count) | PSNR: 28.49 -> 28.88 | Improved |
| use symplectic FlowPET instead of dissipative solver on lesion phantoms | FlowPET · Spherical Lesion Phantom / Synthetic Lesion BrainWeb | SUV ratio | Improved |
Closest papers by meaning
- FourierPET: Deep Fourier-based Unrolled Network for Low-count PET Reconstruction (AAAI 2026 · similarity 0.86)
- BuresTomFlow: Bures-Geometric Flow Matching for Posterior Quantum Tomography (NeurIPS 2026 · similarity 0.84)
- STREAM: Stochastic Riemannian Flow Matching with Anisotropic Decoder for Digital Histopathology Image Generation (NeurIPS 2026 · similarity 0.84)
- Provably Robust Score-Based Diffusion Posterior Sampling for Plug-and-Play Image Reconstruction (NeurIPS 2024 · similarity 0.82)
- Towards Prospective Medical Image Reconstruction via Knowledge-Informed Dynamic Optimal Transport (NeurIPS 2025 · similarity 0.81)
- Physics-Constrained Flow Matching: Sampling Generative Models with Hard Constraints (NeurIPS 2025 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
