Method

SubCat with 20 orientations, 48 pixels height, LDCF features [SubCat48LDCF]
[Anonymous Submission]

Submitted on 25 Sep. 2016 03:20 by
[Anonymous Submission]

Running time:5 s
Environment:1 core @ 2.5 Ghz (Matlab + C/C++)

Method Description:
SubCat with 20 AdaBoost detectors 1 for each view
of the car. Locally Decorrelated Features (P.Dollar
NIPS 2014. Depth 2 trees 2048 trees maximum per
detector.
Parameters:
TBD
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) 74.14 % 61.57 % 48.18 %
Car (Orientation) 28.56 % 24.27 % 19.02 %
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




eXTReMe Tracker