Reproduce “ARIDA - Active railway infrastructure dataset annotation using neural network predictions” 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

Maximilian Dornik, Christian Sieberichs, Simon Geerkens, Alexander Braun

ARIDA converts instance-segmentation predictions into paired track-boundary lines, accepts them when confidence and geometric-consistency checks pass, and prioritizes unresolved images for simulated manual labeling. On sequential OSDaR23 images, a run starting with 50 labeled reached about 79% segmentation mAP, matching the 100-label supervised baseline at roughly 60 labels. Tests covered one class and annotation primitive; savings count labels, not time.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.