Reproduce “Dense Center-Direction Regression for Object Counting and Localization with Point Supervision” 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
Domen Tabernik, Jon Muhovič, Danijel Skočaj
CeDiRNet performs point-supervised object counting and localization by training pixels to predict directions toward their nearest annotated center, then using a separate localization network to identify centers. On Sorghum, it reaches 96.44% localization F1 at a 20-pixel tolerance and count MAE of 1.78. Evaluations cover six remote-sensing datasets, not all image domains or every crowded scene.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.