Tag
Lucida proposes a method for composable indoor scene reconstruction that distributes requirements across parsing, generation, and placement using a VLM policy to create high-fidelity editable replicas from cluttered captures.