Method

Struc info fusion I [Struc info fusion I]


Submitted on 5 Aug. 2020 04:58 by
Liang Junxiong (Hua Zhong University fo science and technology)

Running time:0.05 s
Environment:1 core @ 2.5 Ghz (Python)

Method Description:
Multi-Sensor Fusion for Single
Stage 3D Object Detection
Parameters:
None
Latex Bibtex:
@article{sif,

title={Struc info fusion},

author={Pei An and Junxiong Liang and Jie Ma and Kun Yu and Bin Fang},

journal={Submitted to CVIU},

year={2021},

}

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) 96.59 % 93.31 % 88.23 %
Car (Orientation) 96.24 % 92.71 % 87.55 %
Car (3D Detection) 88.70 % 79.49 % 74.25 %
Car (Bird's Eye View) 94.91 % 89.38 % 84.29 %
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