The KITTI Vision Benchmark Suite

Method

Efficient Stereo Geometry Network for Fast 3D Object Detection [st] [ESGN]

Submitted on 14 Nov. 2021 04:33 by
Aqi Gao (tianjin university)

Running time:		0.06 s
Environment:		GPU @ 2.5 Ghz (Python + C/C++)

Method Description:

Efficient Geometry Feature Network for Fast Stereo
3D Object Detection (IEEE Transactions on Circuits
and Systems for Video Technology)

Parameters:

\alpha=0.2

Latex Bibtex:

@ARTICLE{9869894,
author={Gao, Aqi and Pang, Yanwei and Nie, Jing
and Shao, Zhuang and Cao, Jiale and Guo, Yishun
and Li, Xuelong},
journal={IEEE Transactions on Circuits and
Systems for Video Technology},
title={ESGN: Efficient Stereo Geometry Network
for Fast 3D Object Detection},
year={2022},
volume={},
number={},
pages={1-1},
doi={10.1109/TCSVT.2022.3202810}}

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)	93.07 %	80.58 %	70.68 %
Car (Orientation)	92.74 %	79.84 %	69.76 %
Car (3D Detection)	65.80 %	46.39 %	38.42 %
Car (Bird's Eye View)	78.10 %	58.12 %	49.28 %
Pedestrian (Detection)	44.09 %	32.60 %	29.10 %
Pedestrian (Orientation)	26.02 %	19.17 %	16.90 %
Pedestrian (3D Detection)	14.05 %	10.27 %	9.02 %
Pedestrian (Bird's Eye View)	17.94 %	13.03 %	11.54 %
Cyclist (Detection)	21.13 %	13.45 %	11.72 %
Cyclist (Orientation)	12.50 %	7.73 %	6.80 %
Cyclist (3D Detection)	13.84 %	7.69 %	6.75 %
Cyclist (Bird's Eye View)	15.78 %	9.02 %	7.96 %