Publication detail

"Local Rank Differences" Image Feature Implemented on GPU

POLOK, L. HEROUT, A. ZEMČÍK, P. HRADIŠ, M. JURÁNEK, R. JOŠTH, R.

Original Title

"Local Rank Differences" Image Feature Implemented on GPU

English Title

"Local Rank Differences" Image Feature Implemented on GPU

Type

conference paper

Language

en

Original Abstract

A currently popular trend in object detection and pattern recognition is usage of statistical classifiers, namely AdaBoost and its modifications. The speed performance of these classifiers largely depends on the low level image features they are using: both on the amount of information the feature provides and the executional time of its evaluation. Local Rank Differences is an image feature that is alternative to commonly used haar wavelets. It is suitable for implementation in programmable (FPGA) or specialized (ASIC) hardware, but --- as this paper shows --- it performs very well on graphics hardware (GPU) as well. The paper discusses the LRD features and their properties, describes an experimental implementation of LRD in graphics hardware, presents its empirical performance measures compared to alternative approaches and suggests several notes on practical usage of LRD and proposes directions for future work.

English abstract

A currently popular trend in object detection and pattern recognition is usage of statistical classifiers, namely AdaBoost and its modifications. The speed performance of these classifiers largely depends on the low level image features they are using: both on the amount of information the feature provides and the executional time of its evaluation. Local Rank Differences is an image feature that is alternative to commonly used haar wavelets. It is suitable for implementation in programmable (FPGA) or specialized (ASIC) hardware, but --- as this paper shows --- it performs very well on graphics hardware (GPU) as well. The paper discusses the LRD features and their properties, describes an experimental implementation of LRD in graphics hardware, presents its empirical performance measures compared to alternative approaches and suggests several notes on practical usage of LRD and proposes directions for future work.

Keywords

Local Rank Differences, Object Detection, AdaBoost, Real-Time Detection

RIV year

2008

Released

24.10.2008

Publisher

Springer Verlag

Location

Berlin, Heidelberg

ISBN

978-3-540-88457-6

Book

Proceedings of the 10th International Conference on Advanced Concepts for Intelligent Vision Systems

Edition

Lecture Notes In Computer Science; Vol. 5259

Edition number

NEUVEDEN

Pages from

170

Pages to

181

Pages count

12

Documents

BibTex


@inproceedings{BUT29656,
  author="Lukáš {Polok} and Adam {Herout} and Pavel {Zemčík} and Michal {Hradiš} and Roman {Juránek} and Radovan {Jošth}",
  title=""Local Rank Differences" Image Feature Implemented on GPU",
  annote="A currently popular trend in object detection and pattern recognition is usage of
statistical classifiers, namely AdaBoost and its modifications. The speed
performance of these classifiers largely depends on the low level image features
they are using: both on the amount of information the feature provides and the
executional time of its evaluation. Local Rank Differences is an image feature
that is alternative to commonly used haar wavelets. It is suitable for
implementation in programmable (FPGA) or specialized (ASIC) hardware, but --- as
this paper shows --- it performs very well on graphics hardware (GPU) as well.
The paper discusses the LRD features and their properties, describes an
experimental implementation of LRD in graphics hardware, presents its empirical
performance measures compared to alternative approaches and suggests several
notes on practical usage of LRD and proposes directions for future work.",
  address="Springer Verlag",
  booktitle="Proceedings of the 10th International Conference on Advanced Concepts for Intelligent Vision Systems",
  chapter="29656",
  edition="Lecture Notes In Computer Science; Vol. 5259",
  howpublished="print",
  institution="Springer Verlag",
  year="2008",
  month="october",
  pages="170--181",
  publisher="Springer Verlag",
  type="conference paper"
}