Rank 1st in the leaderboard of SemanticKITTI semantic segmentation (both single-scan and multi-scan) (Nov. 2020) (CVPR2021 Oral)
Geek Repo:Geek Repo
Github PK Tool:Github PK Tool
chky1997 opened this issue 2 years ago · comments
Hi, I notice that local coordinates of voxels and reflectivity are usually smaller than 1 in the n*9 point features, while the raw x and y range from 0 to 80. Is it necessary to do some data normalization?