Method

NeurOCS: Neural NOCS Supervision for Monocular 3D Object Localization [NeurOCS]


Submitted on 24 Jan. 2023 05:47 by
Bingbing Zhuang (NEC)

Running time:0.1 s
Environment:GPU @ 2.5 Ghz (Python)

Method Description:
NeurOCS: Neural NOCS Supervision for Monocular 3D
Object Localization
Parameters:
N/A
Latex Bibtex:
@InProceedings{Min_2023_CVPR,
author = {Min, Zhixiang and Zhuang, Bingbing
and Schulter, Samuel and Liu, Buyu and Dunn,
Enrique and Chandraker, Manmohan},
title = {NeurOCS: Neural NOCS Supervision
for Monocular 3D Object Localization},
booktitle = {CVPR},
month = {June},
year = {2023},
pages = {21404-21414}
}

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) 96.39 % 91.08 % 81.20 %
Car (Orientation) 96.15 % 90.66 % 80.64 %
Car (3D Detection) 29.89 % 18.94 % 15.90 %
Car (Bird's Eye View) 37.27 % 24.49 % 20.89 %
This table as LaTeX

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



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



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



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




eXTReMe Tracker