Method

Part-aware Local and Nonlocal 3D Object Single Stage Detector [la] [PLNL-3DSSD]
[Anonymous Submission]

Submitted on 4 Jan. 2021 06:52 by
[Anonymous Submission]

Running time:0.08 s
Environment:GPU @ 2.5 Ghz (Python + C/C++)

Method Description:
In this work, We attempt to construct the neighbor
points’ relation and different instances’
relation. The local points’ relation module is
utilized for passing message in a surround area
which could enhance the feature representation.
The instances’ relation module is using for
exchanging discriminative information which could
learn instance-level correlation and precise
representation for classification. In order to
further utilize the enhanced feature for promoting
the ability of location of network, we apply 3D
IoU (Intersection over Union) Loss to assist
relation module generating better feature
representation. Anther contribution of our work is
using density for object part information
aggregation to generate stable local
representation.
Parameters:
base_lr=0.002
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) 96.37 % 95.38 % 90.31 %
Car (3D Detection) 88.98 % 81.69 % 74.90 %
Car (Bird's Eye View) 93.00 % 89.36 % 84.18 %
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




eXTReMe Tracker