Method

Sparse Point-Voxel-BEV Single Shot Detection [SPVB-SSD]


Submitted on 7 Dec. 2021 07:40 by
liu xin (personaly)

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

Method Description:
Sparse 3d conv to extract 3d BEV feature from voxel.
Windows self-attention to proposal 3d bbox.
Parameters:
.....
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) 95.80 % 94.49 % 91.90 %
Car (Orientation) 95.78 % 94.31 % 91.61 %
Car (3D Detection) 86.99 % 80.68 % 76.23 %
Car (Bird's Eye View) 91.82 % 88.23 % 85.46 %
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