Face recognition based on Extended Symmetric Local Graph Structure

Andhik Ampuh Yunanto, Darlis Herumurti

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

1 Citation (Scopus)

Abstract

Face recognition is an important area in biometrics and computer vision. A lot of feature extraction can handle face recognition method such as checking pixel neighbor. Local Binary Pattern, Local Graph Structure, and Symmetric Local Graph Structure are an operator of the feature extraction. This research called Extended Symmetric Local Graph Structure which it is an improvement operator from SLGS to build more symmetric neighbor. The result of ESLGS has average accuracy 84.24% in one until five retrieval similarity of YALE dataset image and 80.59% in one until five retrieval similarity of ORL dataset image. The conclusion indicates that our proposed operator has more accuracy than LBP, LGS and SLGS operator. Advantage of proposed method is to provide better performance in accuracy and complexity than other operator.

Original languageEnglish
Title of host publicationProceedings of 2016 International Conference on Information and Communication Technology and Systems, ICTS 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages80-84
Number of pages5
ISBN (Electronic)9781509013791
DOIs
Publication statusPublished - 24 Apr 2017
Event2016 International Conference on Information and Communication Technology and Systems, ICTS 2016 - Surabaya, Indonesia
Duration: 12 Oct 2016 → …

Publication series

NameProceedings of 2016 International Conference on Information and Communication Technology and Systems, ICTS 2016

Conference

Conference2016 International Conference on Information and Communication Technology and Systems, ICTS 2016
Country/TerritoryIndonesia
CitySurabaya
Period12/10/16 → …

Keywords

  • Face Recognition
  • Image Retrieval
  • Symmetric Local Graph Structure

Fingerprint

Dive into the research topics of 'Face recognition based on Extended Symmetric Local Graph Structure'. Together they form a unique fingerprint.

Cite this