Method

mypointnet++ [MP]


Submitted on 17 Oct. 2019 11:09 by
MIAOmiao hen (xmu)

Running time:0.2 s
Environment:1 core @ 2.5 Ghz (C/C++)

Method Description:
Using pointnet++ to perform global feature
extraction and then use sparse convolution to
obtain corresponding features, and finally use
their own designed RPN prediction.
Parameters:
alpha: 0.25
gamma: 2.0
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) 93.86 % 90.50 % 85.17 %
Car (Orientation) 0.63 % 1.51 % 2.03 %
Car (3D Detection) 84.00 % 73.32 % 67.82 %
Car (Bird's Eye View) 90.24 % 86.16 % 78.86 %
Pedestrian (Detection) 60.26 % 48.73 % 45.05 %
Pedestrian (Orientation) 6.41 % 5.39 % 5.14 %
Pedestrian (3D Detection) 43.04 % 33.89 % 31.46 %
Pedestrian (Bird's Eye View) 47.59 % 38.77 % 35.50 %
Cyclist (Detection) 85.05 % 69.52 % 63.17 %
Cyclist (Orientation) 0.62 % 0.97 % 0.89 %
Cyclist (3D Detection) 72.99 % 55.36 % 49.36 %
Cyclist (Bird's Eye View) 77.57 % 60.16 % 54.01 %
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