Hidden Markov models & principal component analysis for multispectral palmprint identification

Abdallah Meraoumia, Maarouf Korichi, Salim Chitroub, Ahmed Bouridane

Research output: Contribution to conferencePaperpeer-review

Abstract

Automatic personal identification from their physical and behavioral traits, called biometrics technologies, is now needed in many fields such as: surveillance systems, access control systems, physical buildings and many more applications. In this paper, we propose an efficient online personal identification system based on Multi-Spectral Palmprint images (MSP) using Hidden Markov Model (HMM) and Principal Components Analysis (PCA). In this study, the band image {RED, BLUE, GREEN and Nearest-InfraRed (NIR)} is rotated with different orientations then applying the PCA technique to each oriented image, to decorrelate the image columns, and concentrate the information content on the first components of the transformed vectors. Thus, the observation vector is formed by concatenate some components of the transformed vectors for all orientations. Subsequently, we use the HMM for modeling the observation vector of each MSP. Finally, log-likelihood scores are used for MSP matching. Our experimental results show the effectiveness and reliability of the proposed approach, which brings both high identification and accuracy rate.
Original languageEnglish
Publication statusPublished - 21 Dec 2015
Event5th International Conference on Information and Communication Technology and Accessbility (ICTA) - Marrakech
Duration: 21 Dec 2015 → …

Conference

Conference5th International Conference on Information and Communication Technology and Accessbility (ICTA)
Period21/12/15 → …

Fingerprint Dive into the research topics of 'Hidden Markov models & principal component analysis for multispectral palmprint identification'. Together they form a unique fingerprint.

Cite this