Method

Efficient Point-based Single Scale 3D Object Detection from Traffic Scenes [SS-3DSSD]
no

Submitted on 11 May. 2023 08:50 by
tang Qiao (XJTU)

Running time:0.014s
Environment:1 core @ 2.5 Ghz (C/C++)

Method Description:
Our method eliminates
the time-consuming multi-scale feature extraction
module used in Point-Net++ and adopts an efficient
single-scale feature extraction method based on
neighbor-attention, significantly improving the
model’s inference speed. Additionally, we
introduce a learning-based sampling method
to overcome the limited receptive fields of
single-scale methods and a multi-level context
feature aggregation module to meet varying feature
requirements at different leve.
Parameters:
no
Latex Bibtex:
no

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) 96.31 % 95.47 % 90.55 %
Car (Orientation) 96.30 % 95.39 % 90.43 %
Car (3D Detection) 87.98 % 81.35 % 76.43 %
Car (Bird's Eye View) 92.61 % 89.27 % 86.10 %
This table as LaTeX

2D object detection results.
This figure as: png eps txt gnuplot



Orientation estimation results.
This figure as: png eps txt gnuplot



3D object detection results.
This figure as: png eps txt gnuplot



Bird's eye view results.
This figure as: png eps txt gnuplot




eXTReMe Tracker