Method

Faster-rcnn+fpn [detectron]
[Anonymous Submission]

Submitted on 9 Sep. 2018 14:58 by
[Anonymous Submission]

Running time:0.01 s
Environment:1 core @ 2.5 Ghz (C/C++)

Method Description:
this is the Fair's detectron. it is used in caffe2.i will do
more for better grade.
Parameters:
max_iter=70000
base_lr=0.001
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) 88.14 % 78.96 % 69.74 %
Car (Orientation) 35.58 % 31.71 % 28.18 %
Pedestrian (Detection) 69.89 % 52.42 % 51.70 %
Pedestrian (Orientation) 41.08 % 31.20 % 30.78 %
Cyclist (Detection) 64.73 % 48.06 % 40.75 %
Cyclist (Orientation) 27.44 % 26.36 % 23.20 %
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