3DSAC: Size Adaptive Clustering for 3D object detection in point clouds

Open Webpage

Hang Yu, Jinhe Su, Guorong Cai, Yingchao Piao, Niansheng Liu, Min Huang

Published: 01 Jan 2023, Last Modified: 06 Nov 2025Int. J. Appl. Earth Obs. Geoinformation 2023EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Refine voting object centers by segmenting the point clouds into foreground and background points.•A size-adaptive feature grouping and implicit semantic object relations for 3D object detection.•Experiments based on ScanNet and SUN RGB-D show that 3DSAD can efficiently detect objects.