← back to paper
arxiv: 2605.31597 · 2 revisions
SOCO: Benchmarking Semantic Object Correspondence in Vision Foundation Models