Research Aim

research aim..

CT / MRI

CT scans are usually acquired with large slice thickness, leading to anisotropic volumes with inconsistent resolution. Such inconsistency hinders diagnosis, motivating research on deep learning-based super-resolution. Existing methods focus on single-slice super-resolution or slice synthesis but rarely leverage the 3D anisotropic property. This work proposes a cross-view texture transfer framework that transfers high-resolution in-plane textures to low-resolution through-plane slices. A multi-reference non-local attention module enhances detail reconstruction, achieving superior performance on public CT datasets.

🔗 Quick Links: [Google Scholar] [LinkedIn] [ResearchGate] [GitHub]