Reproduce “Every Pixel Is an Option: A System One Model for Medical Ultrasound Biometry” 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

Moein Heidari, Junbo Rao, Yudan Chen, Yasamin Medghalchi, River Jiang, Wenjin Chen, David J. Foran, Jasmine Grewal and 1 more

Kev-Bio turns ultrasound biometry questions into searches for caliper endpoints defining a measurement. It uses language-model decision states to query a shared dense image-feature map, producing spatial probability maps from which endpoint coordinates and measurements are derived. On five-fold cross-validation of the FU/GU Biometry 2026 training set, it achieves 22.50-pixel macro-mean landmark error, but unseen tasks and clinical deployment remain untested.

Summary by alphaXiv.

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