Publication detail

Real-time object detection on CUDA

HEROUT, A. JOŠTH, R. JURÁNEK, R. HAVEL, J. HRADIŠ, M. ZEMČÍK, P.

Original Title

Real-time object detection on CUDA

English Title

Real-time object detection on CUDA

Type

journal article - other

Language

en

Original Abstract

The aim of the research described in this article is to accelerate object detection in images and video sequences using graphics processors. It includes algorithmic modifications and adjustments of existing detectors, constructing variants of efficient implementations and evaluation comparing with efficient implementations on the CPUs. This article focuses on detection by statistical classifiers based on boosting. The implementation and the necessary algorithmic alterations are described, followed by experimental measurements of the created object detector and discussion of the results. The final solution outperforms the reference efficient CPU/SSE implementation, by approximately 6-89 for high-resolution videos using nVidia GeForce 9800GTX and Intel Core2 Duo E8200.

English abstract

The aim of the research described in this article is to accelerate object detection in images and video sequences using graphics processors. It includes algorithmic modifications and adjustments of existing detectors, constructing variants of efficient implementations and evaluation comparing with efficient implementations on the CPUs. This article focuses on detection by statistical classifiers based on boosting. The implementation and the necessary algorithmic alterations are described, followed by experimental measurements of the created object detector and discussion of the results. The final solution outperforms the reference efficient CPU/SSE implementation, by approximately 6-89 for high-resolution videos using nVidia GeForce 9800GTX and Intel Core2 Duo E8200.

Keywords

Object detection, Pattern recognition, Acceleration, CUDA, AdaBoost

RIV year

2011

Released

12.08.2011

Publisher

NEUVEDEN

Location

NEUVEDEN

ISBN

1861-8200

Periodical

Journal of Real-Time Image Processing

Year of study

2011

Number

3

State

DE

Pages from

159

Pages to

170

Pages count

12

URL

Documents

BibTex


@article{BUT50548,
  author="Adam {Herout} and Radovan {Jošth} and Roman {Juránek} and Jiří {Havel} and Michal {Hradiš} and Pavel {Zemčík}",
  title="Real-time object detection on CUDA",
  annote="The aim of the research described in this article is to accelerate object
detection in images and video sequences using graphics processors. It includes
algorithmic modifications and adjustments of existing detectors, constructing
variants of efficient implementations and evaluation comparing with efficient
implementations on the CPUs. This article focuses on detection by statistical
classifiers based on boosting. The implementation and the necessary algorithmic
alterations are described, followed by experimental measurements of the created
object detector and discussion of the results. The final solution outperforms the
reference efficient CPU/SSE implementation, by approximately 6-89 for
high-resolution videos using nVidia GeForce 9800GTX and Intel Core2 Duo E8200.",
  address="NEUVEDEN",
  chapter="50548",
  edition="NEUVEDEN",
  howpublished="print",
  institution="NEUVEDEN",
  number="3",
  volume="2011",
  year="2011",
  month="august",
  pages="159--170",
  publisher="NEUVEDEN",
  type="journal article - other"
}