Method

[la]Multi Object Det [MODet]


Submitted on 15 Apr. 2019 23:25 by
o nly ( )

Running time:0.05 s
Environment:GTX1080Ti

Method Description:
XYZ of LIDAR only.
Parameters:
Latex Bibtex:
@inproceedings{zhang2019accurate,
title={Accurate and Real-Time Object
Detection Based on Bird's Eye View on 3D Point
Clouds},
author={Zhang, Yi and Xiang, Zhiyu and Qiao,
Chengyu and Chen, Shuya},
booktitle={2019 International Conference on
3D Vision (3DV)},
pages={214--221},
year={2019},
organization={IEEE}
}

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) 66.06 % 62.54 % 60.04 %
Car (Bird's Eye View) 90.80 % 87.56 % 82.69 %
This table as LaTeX


2D 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