Implementation of face selective attention model on an embedded system

Bumhwi Kim, Hyung Min Son, Yun Jung Lee, Minho Lee

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

This paper proposes a new embedded system which can selectively detect human faces with fast speed. The embedded system is developed by using OMAP 3530 application processor which has DSP and ARM core. Since the embedded system has the limited performance of CPU and memory, we propose a hybrid system combined the YCbCr based bottom-up selective attention with the conventional Adaboost algorithm. The proposed method using the bottom-up selective attention model can reduce not only the false positive error ratio of the Adaboost based face detection algorithm but also the time complexity by finding the candidate regions of the foreground and reducing the regions of interest (ROI) in the image. The experimental results show that the implemented embedded system can successfully work for localizing human faces in real time.

Original languageEnglish
Title of host publicationNeural Information Processing - 19th International Conference, ICONIP 2012, Proceedings
Pages188-195
Number of pages8
EditionPART 5
DOIs
StatePublished - 2012
Event19th International Conference on Neural Information Processing, ICONIP 2012 - Doha, Qatar
Duration: 12 Nov 201215 Nov 2012

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 5
Volume7667 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference19th International Conference on Neural Information Processing, ICONIP 2012
Country/TerritoryQatar
CityDoha
Period12/11/1215/11/12

Keywords

  • Adaboost
  • bottom-up selective attention
  • embedded
  • face detection
  • modified census transform

Fingerprint

Dive into the research topics of 'Implementation of face selective attention model on an embedded system'. Together they form a unique fingerprint.

Cite this