Method

[la]Licar [Licar]


Submitted on 5 May. 2018 17:34 by
qiankun tang (The Institute of Computing Technology of the Chinese Academy of Sciences)

Running time:0.09 s
Environment:GPU @ 2.0 Ghz (Python)

Method Description:
Parameters:
weight decay: 0.0005
momentum: 0.9
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) 41.60 % 33.89 % 35.17 %
Car (Orientation) 18.24 % 15.58 % 16.15 %
Car (3D Detection) 16.25 % 12.88 % 13.67 %
Car (Bird's Eye View) 45.81 % 40.40 % 37.09 %
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



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



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




eXTReMe Tracker