Discriminant spectral hashing for compact palmprint representation

Ying Cong Chen, Meng Hui LIM, Pong Chi YUEN, Jian Huang Lai*

*Corresponding author for this work

Research output: Chapter in book/report/conference proceedingConference contributionpeer-review

7 Citations (Scopus)


When palmprint recognition needs to be run in the device with low processing and small storage capacities, binary representation with low storage overhead, high matching speed and high discrimination power is preferred. However, existing feature extraction methods focus more on matching accuracy than representation compactness, which would result in high storage and operation cost. Inspired by Spectral Hashing that is known for compact-binary-representation extraction in the image retrieval domain, we propose a compact binary feature extraction method called Discriminant Spectral Hashing (DSH). DSH projects the feature to a discriminative subspace and then performs Spectral Hashing to obtain discriminative and compact code. Experiment results on a benchmark palmprint database show that our algorithm outperforms the existing coding-based methods in recognition accuracy with shorter code.

Original languageEnglish
Title of host publicationBiometric Recognition - 8th Chinese Conference, CCBR 2013, Proceedings
Number of pages8
Publication statusPublished - 2013
Event2012 International Conference on Service-Oriented Computing, ICSOC 2012 - Jinan, China
Duration: 16 Nov 201317 Nov 2013

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8232 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference2012 International Conference on Service-Oriented Computing, ICSOC 2012

Scopus Subject Areas

  • Theoretical Computer Science
  • Computer Science(all)

User-Defined Keywords

  • Compact palmprint representation
  • Discriminant spectral hashing
  • Palmprint recognition


Dive into the research topics of 'Discriminant spectral hashing for compact palmprint representation'. Together they form a unique fingerprint.

Cite this