Reproduce “Learning from high-resolution 2D images to reconstruct 3D microstructures” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Wenbo Sui, Lars P. Mikkelsen, Jens W. Andreasen, Daniel Lichau, Harold Phelippeau, Philip J. Withers
PCRD reconstructs fiber instances in pultruded carbon-fiber composites from micro-CT, using registered SEM images as training-time supervision and micro-CT alone at inference. Its anchor-based similarity distribution transfer matches relationships among spatial locations rather than feature vectors. On a held-out cross-section test, ACSD reaches mAP 0.4419 versus 0.4253 for direct training; 3D validation is qualified, not isotropically high-resolution.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.