Human eye location algorithm based on multi-scale self-quotient image and morphological filtering for multimedia big data

Bin Song, Doo Kwon Baik, Shunxian Zhou

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

In order to reduce the effect on the eyes location caused by the variation of illumination and expression, this paper proposes a human eye location algorithm based on the multi-scale self-quotient image and morphological filtering. Firstly, the multi-scale self-quotient image is used to offset the lighting effects on the face, then the morphological open-close operation will be taken to enhance the local features around the eyes and relevant coefficient is used to roughly position the eyes. At last, the variance projection method will be used to analyze the roughly-positioned areas and binarize them to position accurately the central point of the eye. The experiments on the images from JAFFE Database, Yale B Database and AR database have shown that the proposed algorithm can well position the center of the eye, and it is robust to deal with the changes of illumination and expressions.

Original languageEnglish
Pages (from-to)1-13
Number of pages13
JournalMultimedia Tools and Applications
DOIs
Publication statusAccepted/In press - 2017 Jul 24

Fingerprint

Lighting
Big data
Experiments

Keywords

  • Human Eye Location
  • Morphological Filtering
  • Multi-scale Self-quotient Image
  • Multimedia Big Data
  • Variance Projection

ASJC Scopus subject areas

  • Software
  • Media Technology
  • Hardware and Architecture
  • Computer Networks and Communications

Cite this

@article{31f403cb2bda4625bb1ca4395181e6e8,
title = "Human eye location algorithm based on multi-scale self-quotient image and morphological filtering for multimedia big data",
abstract = "In order to reduce the effect on the eyes location caused by the variation of illumination and expression, this paper proposes a human eye location algorithm based on the multi-scale self-quotient image and morphological filtering. Firstly, the multi-scale self-quotient image is used to offset the lighting effects on the face, then the morphological open-close operation will be taken to enhance the local features around the eyes and relevant coefficient is used to roughly position the eyes. At last, the variance projection method will be used to analyze the roughly-positioned areas and binarize them to position accurately the central point of the eye. The experiments on the images from JAFFE Database, Yale B Database and AR database have shown that the proposed algorithm can well position the center of the eye, and it is robust to deal with the changes of illumination and expressions.",
keywords = "Human Eye Location, Morphological Filtering, Multi-scale Self-quotient Image, Multimedia Big Data, Variance Projection",
author = "Bin Song and Baik, {Doo Kwon} and Shunxian Zhou",
year = "2017",
month = "7",
day = "24",
doi = "10.1007/s11042-017-4945-x",
language = "English",
pages = "1--13",
journal = "Multimedia Tools and Applications",
issn = "1380-7501",
publisher = "Springer Netherlands",

}

TY - JOUR

T1 - Human eye location algorithm based on multi-scale self-quotient image and morphological filtering for multimedia big data

AU - Song, Bin

AU - Baik, Doo Kwon

AU - Zhou, Shunxian

PY - 2017/7/24

Y1 - 2017/7/24

N2 - In order to reduce the effect on the eyes location caused by the variation of illumination and expression, this paper proposes a human eye location algorithm based on the multi-scale self-quotient image and morphological filtering. Firstly, the multi-scale self-quotient image is used to offset the lighting effects on the face, then the morphological open-close operation will be taken to enhance the local features around the eyes and relevant coefficient is used to roughly position the eyes. At last, the variance projection method will be used to analyze the roughly-positioned areas and binarize them to position accurately the central point of the eye. The experiments on the images from JAFFE Database, Yale B Database and AR database have shown that the proposed algorithm can well position the center of the eye, and it is robust to deal with the changes of illumination and expressions.

AB - In order to reduce the effect on the eyes location caused by the variation of illumination and expression, this paper proposes a human eye location algorithm based on the multi-scale self-quotient image and morphological filtering. Firstly, the multi-scale self-quotient image is used to offset the lighting effects on the face, then the morphological open-close operation will be taken to enhance the local features around the eyes and relevant coefficient is used to roughly position the eyes. At last, the variance projection method will be used to analyze the roughly-positioned areas and binarize them to position accurately the central point of the eye. The experiments on the images from JAFFE Database, Yale B Database and AR database have shown that the proposed algorithm can well position the center of the eye, and it is robust to deal with the changes of illumination and expressions.

KW - Human Eye Location

KW - Morphological Filtering

KW - Multi-scale Self-quotient Image

KW - Multimedia Big Data

KW - Variance Projection

UR - http://www.scopus.com/inward/record.url?scp=85025591803&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85025591803&partnerID=8YFLogxK

U2 - 10.1007/s11042-017-4945-x

DO - 10.1007/s11042-017-4945-x

M3 - Article

AN - SCOPUS:85025591803

SP - 1

EP - 13

JO - Multimedia Tools and Applications

JF - Multimedia Tools and Applications

SN - 1380-7501

ER -