HDDPM: Heteroscedastic Denoising Diffusion Probabilistic Model for Quantitative Low-Count Brain PET Recovery
arXiv:2606.28513v1 Announce Type: cross Abstract: Positron emission tomography (PET) seeks to balance diagnostic quality with ra-diation dose. Low-count PET noise is non-Gaussian, non-stationary, and spatial-ly dependent. It scales directly with local activity and is shaped by iterative recon-struction and physical corrections. Standard denoising diffusion probabilistic models (DDPMs) ignore these PET properties. Their forward process adds iso-tropic, homoscedastic Gaussian noise to the target. Such an approach fails to cap-ture the realistic physical degradation generated by the imaging syste
The continuous evolution of AI in medical imaging, particularly diffusion probabilistic models, is driving innovation in diagnostic quality and radiation dose management.
Improving PET imaging quality with reduced radiation exposure has significant implications for patient safety, diagnostic accuracy, and the broader healthcare system.
This research suggests a more effective way to process low-count PET data, potentially leading to safer and more accurate medical diagnostics through specialized AI models.
- · Medical AI researchers
- · Healthcare providers
- · Patients undergoing PET scans
- · Medical imaging equipment manufacturers
- · Legacy image reconstruction methods
- · Non-specialized AI models in medical imaging
More precise and safer medical diagnostics become achievable through enhanced low-count PET recovery.
The widespread adoption of specialized diffusion models could accelerate AI integration across other complex medical imaging modalities.
Reduced radiation exposure could lead to increased frequency of preventative or diagnostic PET scans, if costs permit.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI