Method

A new method base on monocular image [NL_M3D]


Submitted on 13 May. 2020 08:17 by
Shujie Luo (Zhejiang University)

Running time:0.2 s
Environment:1 core @ 2.5 Ghz (Python)

Method Description:
A new 3D object detection method base on monocular
images.
Parameters:
\alpha=0.2
Latex Bibtex:

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) 91.31 % 86.80 % 72.37 %
Car (Orientation) 90.88 % 85.32 % 70.87 %
Car (3D Detection) 17.54 % 11.46 % 8.98 %
Car (Bird's Eye View) 24.15 % 15.93 % 12.11 %
Pedestrian (Detection) 58.46 % 45.03 % 39.22 %
Pedestrian (Orientation) 46.64 % 35.20 % 30.56 %
Pedestrian (3D Detection) 5.16 % 3.87 % 3.08 %
Pedestrian (Bird's Eye View) 6.20 % 4.66 % 3.99 %
Cyclist (Detection) 71.09 % 53.51 % 47.07 %
Cyclist (Orientation) 57.44 % 41.19 % 36.24 %
Cyclist (3D Detection) 2.10 % 1.51 % 1.58 %
Cyclist (Bird's Eye View) 2.70 % 2.01 % 1.75 %
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



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



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