PSCT-Net: Geometry-Aware Pediatric Skull CT Reconstruction via Differentiable Back-Projection and Attention-Guided Refinement
arXiv:2606.19867v1 Announce Type: cross Abstract: Computed Tomography (CT) is essential for diagnosing pediatric craniofacial abnormalities, yet poses radiation risks to developing anatomies. Reconstructing 3D CT from sparse bi-planar X-rays offers a low-dose alternative but is severely ill-posed. Existing methods employ geometry-agnostic feature lifting, naively projecting 2D features into 3D without explicit spatial modeling, causing depth ambiguity and degraded osseous boundaries. We present PSCT-Net, a geometry-aware framework with differentiable back-projection. Differentiable back-projec
The increasing computational power and progress in AI, particularly in computer vision and generative models, are enabling more sophisticated medical image reconstruction techniques to address long-standing challenges.
This development offers a significant step towards safer and more accurate diagnostics for vulnerable patient populations, potentially reducing radiation exposure while improving diagnostic precision.
Traditional geometry-agnostic 3D CT reconstruction methods are being superseded by geometry-aware AI frameworks that leverage explicit spatial modeling for enhanced accuracy and reduced artifacting.
- · Pediatric radiology departments
- · Medical AI software developers
- · Patients requiring frequent CT scans
- · Diagnostic imaging equipment manufacturers
- · Manufacturers of older CT reconstruction algorithms
- · Medical facilities relying solely on high-dose traditional CT
Pediatric CT scans become significantly safer due to reduced radiation dosage.
Improved diagnostic accuracy for craniofacial abnormalities leads to earlier and more effective interventions for children.
The methodology could generalize to other forms of medical imaging reconstruction, leading to a broader paradigm shift in diagnostics for various body parts and patient groups.
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