https://arxiv.org/abs/2105.02788
ACORN: Adaptive Coordinate Networks for Neural Scene Representation (Julien N. P. Martel, David B. Lindell, Connor Z. Lin, Eric R. Chan, Marco Monteiro, Gordon Wetzstein)
https://arxiv.org/abs/2105.02788 요즘 보면 정말로 all you need인 것은 positional encoding이 아닌가 싶죠. 전체 공간을 block들로 쪼갠 다음 각 block의 id를 입력으로 받아 그 block 내의 local coordinate에 대한 feature로 매핑하는 네트워크를 사용해서 neural scene representation을 학습합니다. 결과가 무시무시하네요.
#implicit_representation #positional_encoding