Method

MMDetection3D SECOND [MM_SECOND]
TBD

Submitted on 22 Feb. 2024 12:14 by
Ye Wen (SUES)

Running time:0.05 s
Environment:GPU @ >3.5 Ghz (Python)

Method Description:
Reproduce SECOND based on MMDetection3D
Parameters:
mmdetection3d v1.4.0
batch size = 16
lr = 0.00018
Latex Bibtex:
None

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) 95.14 % 91.71 % 86.75 %
Car (Orientation) 94.85 % 90.75 % 85.58 %
Car (3D Detection) 82.02 % 72.68 % 66.27 %
Car (Bird's Eye View) 90.52 % 86.39 % 81.49 %
Pedestrian (Detection) 60.59 % 49.60 % 46.70 %
Pedestrian (Orientation) 54.97 % 44.69 % 41.93 %
Pedestrian (3D Detection) 44.12 % 35.19 % 32.02 %
Pedestrian (Bird's Eye View) 49.46 % 40.22 % 37.46 %
Cyclist (Detection) 85.26 % 73.63 % 67.06 %
Cyclist (Orientation) 84.69 % 72.74 % 65.85 %
Cyclist (3D Detection) 73.41 % 57.81 % 51.68 %
Cyclist (Bird's Eye View) 77.98 % 62.61 % 55.67 %
This table as LaTeX


2D object detection results.
This figure as: png eps txt gnuplot



Orientation estimation results.
This figure as: png eps txt gnuplot



3D object detection results.
This figure as: png eps txt gnuplot



Bird's eye view results.
This figure as: png eps txt gnuplot



2D object detection results.
This figure as: png eps txt gnuplot



Orientation estimation results.
This figure as: png eps txt gnuplot



3D object detection results.
This figure as: png eps txt gnuplot



Bird's eye view results.
This figure as: png eps txt gnuplot



2D object detection results.
This figure as: png eps txt gnuplot



Orientation estimation results.
This figure as: png eps txt gnuplot



3D object detection results.
This figure as: png eps txt gnuplot



Bird's eye view results.
This figure as: png eps txt gnuplot




eXTReMe Tracker