The LLFF dataset contains 8 forward-facing scenes. Following [19, 26, 35, 42], we take every 8-th image as the novel views for testing. The input views are evenly sampled across the remaining views.
BibTex:
Before browse our site, please accept our cookies policy
