Method

A Unified Pipeline for 3D Detection and Velocity Estimation of Vehicles [la] [PC-CNN-V2]


Submitted on 5 Jan. 2019 12:27 by
xinxin du (Venti Technologies)

Running time:0.5 s
Environment:GPU @ 2.5 Ghz (Matlab + C/C++)

Method Description:
A more advanced version than our previous work as
described in the reference. The biggest difference is that it
fuses both deep features from image and point cloud.
Parameters:
N.A.
Latex Bibtex:
@INPROCEEDINGS{8461232,
author={X. Du and M. H. Ang and S. Karaman and D. Rus},
booktitle={2018 IEEE International Conference on Robotics
and Automation (ICRA)},
title={A General Pipeline for 3D Detection of Vehicles},
year={2018},
pages={3194-3200},
doi={10.1109/ICRA.2018.8461232},
ISSN={2577-087X},
month={May},}

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) 96.06 % 95.20 % 89.37 %
Car (3D Detection) 85.57 % 73.79 % 65.65 %
Car (Bird's Eye View) 91.19 % 87.40 % 79.35 %
This table as LaTeX

2D object detection 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