VascuSR: Iterative Anatomy-Aware Diffusion for CT Super-Resolution

Key points

• Generative super-resolution (SR) can recover sharp detail in computed tomography (CT) images from severely downsampled observations, but the missing high-frequency content is not uniquely determined by the low-resolution input.1
• First, to provide richer anatomical guidance, we use an explicit 3D vascular representation that captures structural information beyond individual slices.1
• This representation also simplifies vascular refinement by allowing a lightweight 3D generative refiner to focus on vascular structure rather than complex CT appearance.1
• On re-annotated MSD Task08 and HVM datasets, VascuSR outperforms interpolation and learning-based SR baselines in segmentation accuracy (Dice and clDice) and 3D vascular structure (connectivity and mean centerline distance), while maintaining image quality comparable to its SR backbone.1

Sources (2)

• [1]VascuSR: Iterative Anatomy-Aware Diffusion for CT Super-Resolution
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 14, 05:12 PM

  To overcome this challenge, we introduce VascuSR, an iterative anatomy-aware diffusion framework that combines image reconstruction with explicit 3D vascular guidance.

  Generative super-resolution (SR) can recover sharp detail in computed tomography (CT) images from severely downsampled observations, but the missing high-frequency content is not uniquely determined by the low-resolution input.

• [2]VascuSR: Iterative Anatomy-Aware Diffusion for CT Super-Resolution
  Hugging Face Daily Papers · Oct 14, 12:00 AM · same content

Extractive summary: sentences quoted from the sources.