Palm vein recognition using scale invariant feature transform with RANSAC mismatching removal

Shi Chuan, Soh and M. Z., Ibrahim and Marlina, Yakno and D.J., Mulvaney (2017) Palm vein recognition using scale invariant feature transform with RANSAC mismatching removal. In: IT Convergence and Security 201. Lecture Notes in Electrical Engineering, 449 . Springer Verlag, Berlin, Germany, pp. 1-8. ISBN 978-981106450-0

	Pdf 68. Palm Vein Recognition using Scale Invariant Feature Transform with RANSAC Mismatching Removal.pdf - Published Version Restricted to Repository staff only Download (454kB) | Request a copy
Preview	Pdf 68.1 Palm Vein Recognition using Scale Invariant Feature Transform with RANSAC Mismatching Removal.pdf - Published Version Download (60kB) | Preview

Abstract

Palm vein recognition has getting more attention and popular among all other biometrics methods. In order to apply this type of recognition system to society, obtain an accurate reading robustly and effectively become the most pop-ular research topic in this field. However, there are still an unsolved issues on accurate palm vein recognition although there are several research done. In this paper, impact of Random Sample Consensus (RANSAC) point mismatching re-moval and different wavelength spectrum to the recognition rate will be dis-cussed. CASIA Multi Spectral Palm Print Image database is used for this re-search. Scale Invariant Feature Transform (SIFT) and RANSAC mismatching removal will be adopted for vein extraction and point feature matching with Eu-clidean Distance. The results shows that SIFT algorithm with RANSAC mis-matching point removal achieved better recognition rate than without mismatch-ing point removal technique used. It can be proved that RANSAC mismatching point removal are able to remove outlier with preserving the correct point by im-proving the Equal Error Rate (EER) in recognition systems. In palm vein recog-nition system, higher wavelength spectrum of palm vein image will achieved higher verification rate. This can be shows that vein pattern are able and success-fully extract on the image with higher wavelength spectrum.

Item Type:	Book Chapter
Additional Information:	Index by Scopus
Uncontrolled Keywords:	Vein Recognition; Scale Invariant Feature Transform; Random Sample Consensus
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science T Technology > TK Electrical engineering. Electronics Nuclear engineering
Faculty/Division:	Faculty of Electrical & Electronic Engineering
Depositing User:	Pn. Hazlinda Abd Rahman
Date Deposited:	24 May 2018 07:41
Last Modified:	20 Jul 2018 03:11
URI:	http://umpir.ump.edu.my/id/eprint/20561
Download Statistic:	View Download Statistics

Actions (login required)

View Item