Search papers, labs, and topics across Lattice.
1
0
3
LLMs beat VLMs at judging semantic layout quality, but image-based VLMs are the best critics for correcting orientation and semantics in 3D indoor scene synthesis.