Method

Object Detection on Grid Maps Capturing Uncertainty [la] [TopNet-UncEst]


Submitted on 5 Mar. 2019 10:28 by
Sascha Wirges (Karlsruhe Institute of Technology)

Running time:0.09 s
Environment:NVIDIA GeForce 1080 Ti (tensorflow-gpu)

Method Description:
See paper for details.
Parameters:
See paper for details.
Latex Bibtex:
@misc{wirges2019capturing, title={Capturing
Object Detection Uncertainty in Multi-Layer Grid
Maps}, author={Sascha Wirges and Marcel Reith-
Braun and Martin Lauer and Christoph Stiller},
year={2019}, eprint={1901.11284}, archivePrefix=
{arXiv}, primaryClass={cs.RO}}

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) 7.24 % 6.24 % 5.42 %
Car (3D Detection) 3.24 % 3.02 % 2.26 %
Car (Bird's Eye View) 72.05 % 59.67 % 51.67 %
Pedestrian (Detection) 13.00 % 8.58 % 7.38 %
Pedestrian (3D Detection) 3.42 % 1.87 % 1.73 %
Pedestrian (Bird's Eye View) 6.88 % 4.60 % 3.79 %
Cyclist (Detection) 18.14 % 12.00 % 11.85 %
Cyclist (3D Detection) 7.13 % 4.54 % 3.81 %
Cyclist (Bird's Eye View) 12.31 % 9.18 % 8.14 %
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



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



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