Neural Radiance Fields for Manhattan Scenes with Unknown Manhattan Frame. (arXiv:2212.01331v2 [cs.CV] UPDATED)

Novel view synthesis and 3D modeling using implicit neural field
representation are shown to be very effective for calibrated multi-view
cameras. Such representations are known to benefit from additional geometric
and semantic supervision. Most existing methods that exploit additional
supervision require dense pixel-wise labels or localized scene priors. These
methods cannot benefit from high-level vague scene priors provided in terms of
scenes' descriptions. In this work, we aim to leverage the geometric prior of
Manhattan scenes to improve the implicit …

3d modeling aim arxiv benefit cameras exploit fields labels modeling neural radiance field neural radiance fields novel pixel prior representation semantic supervision synthesis terms work