Method

[la]CIA-SSD [CIA-SSD]
https://github.com/Vegeta2020/CIA-SSD

Submitted on 10 Sep. 2020 14:54 by
Vincent Chen (The Chinese University of Hong Kong)

Running time:0.03 s
Environment:1 core @ 2.5 Ghz (Python + C/C++)

Method Description:
TBD
Parameters:
TBD
Latex Bibtex:
@inproceedings{zheng2020ciassd,
title={CIA-SSD: Confident IoU-Aware Single-Stage
Object Detector From Point Cloud},
author={Zheng, Wu and Tang, Weiliang and Chen,
Sijin and Jiang, Li and Fu, Chi-Wing},
booktitle={AAAI},
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) 96.87 % 93.72 % 86.20 %
Car (Orientation) 96.65 % 93.34 % 85.76 %
Car (3D Detection) 89.59 % 80.28 % 72.87 %
Car (Bird's Eye View) 93.74 % 89.84 % 82.39 %
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