Method

SegVoxelNet [SegVoxelNet]


Submitted on 21 Apr. 2019 12:29 by
Hongwei Yi (Peking University )

Running time:0.04 s
Environment:1 core @ 2.5 Ghz (Python)

Method Description:
https://arxiv.org/abs/2002.05316(accepted by
ICRA2020)
Parameters:
TBD
Latex Bibtex:
@inproceedings{yi2020SegVoxelNet,
title={SegVoxelNet: Exploring Semantic Context
and
Depth-aware Features for 3D Vehicle Detection from
Point Cloud},
author={Yi, Hongwei and Shi, Shaoshuai and Ding,
Mingyu and Sun, Jiankai and Xu, Kui and Zhou, Hui
and Wang, Zhe and Li, Sheng and Wang, Guoping},
booktitle={ICRA},
year={2020}
}

Detailed Results

Object detection and orientation estimation results. Results for object detection are given in terms of average precision (AP) and results for joint object detection and orientation estimation are provided in terms of average orientation similarity (AOS).


Benchmark Easy Moderate Hard
Car (Detection) 96.00 % 92.73 % 87.60 %
Car (Orientation) 95.86 % 92.16 % 86.90 %
Car (3D Detection) 86.04 % 76.13 % 70.76 %
Car (Bird's Eye View) 91.62 % 86.37 % 83.04 %
This table as LaTeX


2D object detection results.
This figure as: png eps pdf txt gnuplot



Orientation estimation results.
This figure as: png eps pdf txt gnuplot



3D object detection results.
This figure as: png eps pdf txt gnuplot



Bird's eye view results.
This figure as: png eps pdf txt gnuplot




eXTReMe Tracker