Reproduce “AgriCountDINO: Parameter-Efficient Exemplar-Guided Counting and Localization in Agriculture” 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

Shengjie Guo, Xin Li, Borjana Arsova, Hanno Scharr, Silvio Salvi

AgriCountDINO lets users specify a target with a few example boxes and predicts a point for each instance, making counts inspectable. A frozen DINOv3-pretrained ConvNeXt-B conditions multiscale features on exemplar appearance and size, while recovery and offset prediction address crowded targets. On TPC-268 validation, three-shot MAE was 11.92 versus 13.20 for TasselNetV4; this result is not universal.

Summary by alphaXiv.

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