STD is a two-stage point-cloud 3D detector using point-seeded spherical anchors for proposals, PointsPool for feature aggregation, and an IoU branch, reporting large gains over prior methods on KITTI especially on hard cases at >10 FPS.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
STD: Sparse-to-Dense 3D Object Detector for Point Cloud
STD is a two-stage point-cloud 3D detector using point-seeded spherical anchors for proposals, PointsPool for feature aggregation, and an IoU branch, reporting large gains over prior methods on KITTI especially on hard cases at >10 FPS.