Method

Multi-View Reprojection Architecture [MVRA + I-FRCNN+]


Submitted on 12 Aug. 2019 09:38 by
Heemin Choi (Samsung Advanced Institute of Technology)

Running time:0.18 s
Environment:GPU @ 2.5 Ghz (Python)

Method Description:
Parameters:
Latex Bibtex:
@InProceedings{Choi_2019_ICCV,
author = {Choi, Hee Min and Kang, Hyoa and Hyun,
Yoonsuk},
title = {Multi-View Reprojection Architecture for
Orientation Estimation},
booktitle = {The IEEE International Conference on
Computer Vision (ICCV) Workshops},
month = {Oct},
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) 95.87 % 94.98 % 82.52 %
Car (Orientation) 95.66 % 94.46 % 81.74 %
Car (3D Detection) 5.19 % 3.27 % 2.49 %
Car (Bird's Eye View) 9.05 % 5.84 % 4.50 %
This table as LaTeX


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



Orientation estimation 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