I am wondering how many ground truth images of different views are generally required to produce a high-quality scene, and if there are possible improvements to reduce it.
I am wondering how many ground truth images of different views are generally required to produce a high-quality scene, and if there are possible improvements to reduce it.