arrow
Return

PointSlice: Accurate and efficient slice-based representation for 3D object detection from point clouds

delete2026-03-11
delete0
PRE
AI
Q
Qifeng Liu
D
Dawei Zhao
Y
Yabo Dong
L
Liang Xiao
王娟 cover
王娟 (Juan Wang)
C
Chen Min
F
Fuyang Li
W
Weizhong Jiang
D
Dongming Lu
Y
Yiming Nie
DOI:10.1016/j.patcog.2026.113449delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
• A novel slice-based encoding method for 3D object detection in outdoor point clouds is proposed. • A Slice Interaction Network (SIN) enables effective information interaction between slices. • State-of-the-art performance (66.7 mAP) is achieved on the nuScenes validation dataset.
Keywords:
3D object detection
point clouds
slice-based encoding
Slice Interaction Network
nuScenes

Journal

Pattern Recognition cover
Pattern Recognition
IF:
7.6
Papers:
1.3W
Citations:
4.5W

Organization

Z
Zhejiang University
Scholars:
1.5W
Papers: 5.2K
Citations: 17.8W
T
Tsinghua University
Scholars:
8.6K
Papers: 4.1K
Citations: 17.7W
D
defense innovation institute
Scholars:
105
Papers: 35
Citations: 0
researcher View more organizations