Fetching the paper…

Efficient Region-Aware Neural Radiance Fields for High-Fidelity Talking Portrait Synthesis · Around