Method

Accurate Monocular 3D Object Detection [AM3D]


Submitted on 18 Mar. 2019 20:45 by
xinzhu ma (Dalian University of Technology)

Running time:0.4 s
Environment:GPU @ 2.5 Ghz (Python + C/C++)

Method Description:
Accurate Monocular 3D Object Detection via Color-
Embedded 3D Reconstruction
Parameters:
r=0.5
Latex Bibtex:
@inproceedings{ma2019accurate,
title={Accurate Monocular Object Detection via Color-
Embedded 3D Reconstruction for Autonomous Driving},
author={X. Ma and Z. Wang and H. Li and P. Zhang and W.
Ouyang and X. Fan},
booktitle={Proceedings of the IEEE international
Conference on Computer Vision (ICCV)},
year={2019}
}

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) 92.55 % 88.71 % 77.78 %
Car (3D Detection) 16.50 % 10.74 % 9.52 %
Car (Bird's Eye View) 25.03 % 17.32 % 14.91 %
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