Search papers, labs, and topics across Lattice.
Adobe Research
2
0
3
Oriented 3D primitives can dramatically enhance the fidelity of multi-agent video authoring, yielding up to 3x better trajectory accuracy compared to traditional methods.
Forget geometric LODs: tokenizing 3D shapes by semantic salience unlocks SOTA reconstruction and efficient autoregressive generation with 10x-1000x fewer tokens.
