Publication result detail

Accelerated Object Detection

HAVEL, J.

Original Title

Accelerated Object Detection

English Title

Accelerated Object Detection

Type

Paper in proceedings outside WoS and Scopus

Original Abstract

This paper presents the improved version of a rapid object detector. This detector uses Waldboost classifier with Local Rank Differences features for face detection. Because of it's computational complexity, original detector was parallelized to take advantage of modern multicore processors. Further speedup was achieved by fixed step scaling with heavy use of SSE instruction set and template metaprogramming.

English abstract

This paper presents the improved version of a rapid object detector. This detector uses Waldboost classifier with Local Rank Differences features for face detection. Because of it's computational complexity, original detector was parallelized to take advantage of modern multicore processors. Further speedup was achieved by fixed step scaling with heavy use of SSE instruction set and template metaprogramming.

Keywords

object detection, classification, adaboost, waldboost, local rank differences, parallelization, OpenMP, SSE, template metaprogramming

Key words in English

object detection, classification, adaboost, waldboost, local rank differences, parallelization, OpenMP, SSE, template metaprogramming

Authors

HAVEL, J.

RIV year

2010

Released

23.04.2009

Publisher

Faculty of Electrical Engineering and Communication BUT

Location

Brno

ISBN

978-80-214-3870-5

Book

Proceedings of the 15th Conference Student EEICT 2009

Edition

Volume 4

Pages from

456

Pages to

460

Pages count

5

BibTex

@inproceedings{BUT30206,
  author="Jiří {Havel}",
  title="Accelerated Object Detection",
  booktitle="Proceedings of the 15th Conference Student EEICT 2009",
  year="2009",
  series="Volume 4",
  pages="456--460",
  publisher="Faculty of Electrical Engineering and Communication BUT",
  address="Brno",
  isbn="978-80-214-3870-5"
}