Method

RFCN [RFCN]
[Anonymous Submission]

Submitted on 25 Sep. 2017 17:58 by
[Anonymous Submission]

Running time:0.3 s
Environment:GPU @ 1.0 Ghz (Python + C/C++)

Method Description:
Testing the performance of RFCN
Parameters:
Using the default ones.
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) 88.69 % 79.44 % 70.06 %
Car (Orientation) 37.75 % 33.59 % 29.82 %
Pedestrian (Detection) 74.44 % 58.06 % 51.14 %
Pedestrian (Orientation) 44.20 % 35.26 % 31.60 %
Cyclist (Detection) 63.26 % 51.19 % 44.54 %
Cyclist (Orientation) 32.01 % 26.92 % 23.86 %
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