TheShadow29/awesome-grounding

Organizes papers across multiple visual grounding subtasks—including referring expressions, video activity localization, 3D grounding, and grounded description—with curated dataset references and implementation links. Papers are chronologically arranged within each task category, enabling researchers to track methodological evolution from foundational work to recent pretraining approaches. Maintains quality through manual review of papers before inclusion and accepts community contributions via pull requests with standardized MLA formatting and metadata requirements.

1,125 stars. No commits in the last 6 months.