Method

Stereo Frustums [st] [la] [SF]


Submitted on 8 Feb. 2020 17:42 by
Xi Mo (University of Kansas)

Running time:0.5 s
Environment:1 core @ 2.5 Ghz (Python)

Method Description:
manuscript in review
Parameters:
manuscript in review
Latex Bibtex:
manuscript in review

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) 60.62 % 46.68 % 38.22 %
Car (3D Detection) 58.88 % 51.92 % 44.59 %
Car (Bird's Eye View) 74.20 % 65.74 % 58.35 %
Pedestrian (Detection) 67.73 % 51.83 % 47.45 %
Pedestrian (3D Detection) 31.61 % 24.84 % 21.96 %
Pedestrian (Bird's Eye View) 37.16 % 29.77 % 26.61 %
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




eXTReMe Tracker