EPC-3D-Diff: Equivariant Physics Consistent Conditional 3D Latent Diffusion for CBCT to CT Synthesis
arXiv:2605.20470v1 Announce Type: cross Abstract: Cone-beam CT (CBCT) is routinely acquired during radiotherapy for patient setup, but its quantitative reliability is degraded by scatter, noise, and reconstruction artifacts, limiting Hounsfield Unit (HU) accuracy. We propose EPC-3D-Diff, a novel conditional 3D latent diffusion framework for volumetric CBCT to CT synthesis that introduces a projection domain equivariance loss derived from acquisition physics. Unlike common image domain equivariance, we exploit the fact that an in plane rotation of the volume corresponds to an angular shift in i
This research is emerging as AI techniques mature and become increasingly integrated into specialized, high-stakes fields like medical imaging for radiation therapy.
Improving the accuracy of medical imaging for radiation therapy directly enhances treatment efficacy and patient outcomes, representing a critical application of advanced AI in healthcare.
This advancement changes the reliability and utility of CBCT scans, mitigating current limitations and potentially broadening their diagnostic and therapeutic applications.
- · Medical imaging companies
- · Radiotherapy equipment manufacturers
- · Hospitals and cancer treatment centers
- · Patients undergoing radiation therapy
- · Developers of less accurate or older CBCT correction methods
- · Diagnostic imaging centers reliant on traditional CT for specific planning steps
Improved radiotherapy planning accuracy and reduced treatment side effects due to better image quality from CBCT.
Increased adoption of CBCT for daily treatment guidance, potentially reducing the need for separate planning CT scans.
The methodology could inspire similar equivariance-based AI improvements in other medical imaging modalities or physics-constrained scientific fields.
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