Method

SD3DOD [SD3DOD]
[Anonymous Submission]

Submitted on 14 Nov. 2021 09:04 by
[Anonymous Submission]

Running time:0.04 s
Environment:GPU @ 2.5 Ghz (Python)

Method Description:
Set distillation method for sparse point clouds
acquired by low-channel LiDAR.
Parameters:
16-channel LiDAR input
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) 92.60 % 81.64 % 75.97 %
Car (Orientation) 92.32 % 80.64 % 74.74 %
Car (3D Detection) 76.09 % 62.00 % 55.46 %
Car (Bird's Eye View) 86.82 % 76.96 % 70.05 %
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