Improvement on gabor texture feature based biometric analysis using image blurring

Da Huang, Kunai Zhang, Dapeng Zhang

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

Abstract

Images without blurring are usually considered as high quality samples in biometric recognition. Image acquisition systems are carefully designed in order to capture clear images. However, experimental results show that the performance of Gabor texture feature based biometric recognition methods can be improved by image blurring. The experiments were conducted on the PolyU Palmprint Database using CompCode as well as on the CASIA Iris Database using Iris Code. The blurring method is to adopt a Gaussian filter to the images during pre-processing. Results indicate that there is an optimal range of each dataset and if all the images are blurred to this range, the performance of the whole dataset will reach optimal. A scheme is also proposed to find the optimal range and to blur an image to this range.
Original languageEnglish
Title of host publicationIntelligence Science and Big Data Engineering
Subtitle of host publicationImage and Video Data Engineering - 5th International Conference, IScIDE 2015, Revised Selected Papers
PublisherSpringer Verlag
Pages420-430
Number of pages11
ISBN (Print)9783319239873
DOIs
Publication statusPublished - 1 Jan 2015
Event5th International Conference on Intelligence Science and Big Data Engineering, IScIDE 2015 - Suzhou, China
Duration: 14 Jun 201516 Jun 2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9242
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference5th International Conference on Intelligence Science and Big Data Engineering, IScIDE 2015
CountryChina
CitySuzhou
Period14/06/1516/06/15

Keywords

  • Gabor filter
  • Gaussian filter
  • Image blurring
  • Iris recognition
  • Palmprint recognition

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this