You are correct. I was confusing this technique with Novel View Synthesis through diffusion (recent paper: https://arxiv.org/abs/2408.06157 https://arxiv.org/ab

You are correct. I was confusing this technique with Novel View Synthesis through diffusion (recent paper: https://arxiv.org/abs/2408.06157 https://arxiv.org/abs/2408.06157) where inference means generating frames rather than points.