The KITTI Vision Benchmark Suite

Method

3D Part-Aware and Aggregation Neural Network for Object Detection from Point Cloud [la] [MMLab-PartA^2]
https://github.com/sshaoshuai/PCDet

Submitted on 8 Dec. 2019 05:50 by
Shaoshuai Shi (CUHK)

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

Method Description:

Results with previous evaluation metric (submitted
on 27 Jun 2019):

Car (3D Detection) 85.94 % 77.86 % 72.00 %
Car (Bird's Eye View) 89.52 % 84.76 % 81.47 %
Cyclist (3D Detection) 78.58 % 62.73 % 57.74 %
Cyclist (Bird's Eye View) 81.91 % 68.12 % 61.92 %

Parameters:

Latex Bibtex:

@article{shi2020part,
title={From Points to Parts: 3D Object Detection from
Point Cloud with Part-aware and Part-aggregation
Network},
author={Shi, Shaoshuai and Wang, Zhe and Shi, Jianping
and Wang, Xiaogang and Li, Hongsheng},
journal={IEEE Transactions on Pattern Analysis and
Machine Intelligence},
publisher={IEEE},
year={2020}
}

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)	95.03 %	91.86 %	89.06 %
Car (Orientation)	95.00 %	91.73 %	88.86 %
Car (3D Detection)	87.81 %	78.49 %	73.51 %
Car (Bird's Eye View)	91.70 %	87.79 %	84.61 %
Pedestrian (Detection)	68.78 %	57.96 %	54.01 %
Pedestrian (Orientation)	63.51 %	52.20 %	48.27 %
Pedestrian (3D Detection)	53.10 %	43.35 %	40.06 %
Pedestrian (Bird's Eye View)	59.04 %	49.81 %	45.92 %
Cyclist (Detection)	88.90 %	78.29 %	71.19 %
Cyclist (Orientation)	88.70 %	77.52 %	70.41 %
Cyclist (3D Detection)	79.17 %	63.52 %	56.93 %
Cyclist (Bird's Eye View)	83.43 %	68.73 %	61.85 %