Method

Binary Integer Programming with Heterogeneous Features [BIP-HETERO]


Submitted on 13 Jul. 2015 12:16 by
Alhayat Ali Mekonnen (LAAS-CNRS)

Running time:~2 s
Environment:1 core @ 2.5 Ghz (C/C++)

Method Description:
This detector is based on heterogeneous pool of
features and binary integer programming (BIP)
based feature selection. The detector employs the
standard cascade of rejectors configuration. But,
instead of using AdaBoost solely for feature
selection and classifier building, it uses an
intermediary feature selection strategy based on
BIP. The BIP is used to select pertinent features
taking both detection performance and computation
time explicitly into consideration. The
heterogeneous feature pool is composed of Haar
like, EOH, LBP, CSS, and HOG features.
Parameters:
Please refer to the associated publication for the
list of parameters.
Latex Bibtex:
@INPROCEEDINGS{Mekonnen2014ICPR,
author={Mekonnen, A.A. and Lerasle, F. and
Herbulot, A. and Briand, C.},
booktitle={Pattern Recognition (ICPR), 2014 22nd
International Conference on},
title={People Detection with Heterogeneous
Features and Explicit Optimization on Computation
Time},
year={2014},
month={Aug},
pages={4322-4327},
}

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
Pedestrian (Detection) 8.51 % 7.05 % 6.30 %
This table as LaTeX


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




eXTReMe Tracker