A distributed weighted voting approach for accurate eye center estimation

Gagandeep Singh, Maheshkumar H. Kolekar

Research output: Contribution to journalArticlepeer-review

Abstract

This paper proposes a novel approach for accurate estimation of eye center in face images. A distributed voting based approach in which every pixel votes is adopted for potential eye center candidates. The votes are distributed over a subset of pixels which lie in a direction which is opposite to gradient direction and the weightage of votes is distributed according to a novel mechanism. First, image is normalized to eliminate illumination variations and its edge map is generated using Canny edge detector. Distributed voting is applied on the edge image to generate different eye center candidates. Morphological closing and local maxima search are used to reduce the number of candidates. A classifier based on spatial and intensity information is used to choose the correct candidates for the locations of eye center. The proposed approach was tested on BioID face database and resulted in better Iris detection rate than the state-of-the-art. The proposed approach is robust against illumination variation, small pose variations, presence of eye glasses and partial occlusion of eyes.

Original languageEnglish (US)
Pages (from-to)292-297
Number of pages6
JournalDefence Science Journal
Volume63
Issue number3
DOIs
StatePublished - May 2013
Externally publishedYes

Keywords

  • Biomarkers
  • Biometrics
  • Eye center estimation
  • Iris recognition

ASJC Scopus subject areas

  • General Chemical Engineering
  • Biomedical Engineering
  • Mechanical Engineering
  • General Physics and Astronomy
  • Computer Science Applications
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'A distributed weighted voting approach for accurate eye center estimation'. Together they form a unique fingerprint.

Cite this