[RSS 2025] ROMAN: a view-invariant global localization method that matches objects from different robot views for reliable pose estimation even when a scene is observed from opposite views
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).