Method

yolov3800 [yolo800]


Submitted on 13 Dec. 2018 12:42 by
Hai Wang (Jiangsu University)

Running time:0.13 s
Environment:4 cores @ 2.5 Ghz (Python)

Method Description:
yolov3 800input
Parameters:
0.025
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) 76.45 % 73.00 % 64.68 %
Car (Orientation) 31.53 % 30.47 % 27.01 %
Pedestrian (Detection) 71.11 % 55.49 % 53.92 %
Pedestrian (Orientation) 40.42 % 31.42 % 30.50 %
Cyclist (Detection) 64.64 % 49.15 % 43.58 %
Cyclist (Orientation) 28.20 % 21.69 % 19.53 %
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



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



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



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



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




eXTReMe Tracker