Paper Title :Multimodal Biometric Authentication System Using A Feature-Level Fusion of Dorsal Hand, Palm and Finger Veins
Author :Yahya Bare Hajon, Mir Omid Mirzada, AozirNoorestani, Kapil Sharma
Article Citation :Yahya Bare Hajon ,Mir Omid Mirzada ,AozirNoorestani ,Kapil Sharma ,
(2023 ) " Multimodal Biometric Authentication System Using A Feature-Level Fusion of Dorsal Hand, Palm and Finger Veins " ,
International Journal of Advances in Science, Engineering and Technology(IJASEAT) ,
pp. 27-31,
Volume-11,Issue-3
Abstract : As technology and innovation have advanced,thedemandformorereliableauthenticationhasgrowninrecent years. In
particular, the use of biometric authenticationhasbecomeincreasinglypopularduetoitsaccuracyandconvenience. The
Traditional biometric authentication systemsdepend on a single physical characteristic, like a fingerprint orface
recognition.However, thesemethods aresusceptible tospoofingand othertypesof
attacks.Tosolvethisproblem,multimodalbiometricauthenticationthatcombinesseveralfeatures has been proposed as
asolution.In thisstudy, weintroduceamultimodalbiometricauthenticationtechniquethat fuses dorsal, finger and palm vein
images at feature levelusing local binary pattern (LBP), principal component analysis(PCA) and k-nearest neighbor (KNN)
classifiers. The proposedsystem extracts the dorsal hand, palm and finger veins imagesfrom a single scan of the user’s hand.
Images are preprocessedtoeliminatethenoiseandenhancefeaturesbeforebeingextractedusingthelocalbinarypatterntechnique.
Theprincipal
component analysis is applied to combine elementsfrom local binary patterns to decrease the size of the featurevector.
Lastly, authentication is achieved by implementing a K-Nearest Neighbor (KNN) classifier. The proposed method
isevaluatedusingtheSDUMLAT-HMTdatabase,theVERAPalm vein database, and the Pontificia Universidad
Javeriana'scollection of Dorsal Hand Vein Images. The proposed systemoutperforms all other existing methods with 98%
accuracy anda minimum EER of 0.01%. The proposed method provides aneffective approach to user authentication. It is
very reliable,accurate and effective. Additionally, it can precisely distinguishbetweenlegitimate users andimposterusers.
Keywords:Biometricauthentication,MultimodalBiometricSystem, dorsal and vein,palm
vein,fingerveins,principalcomponentanalysis(PCA),featureextraction,localbinarypatterns(LBP),fusion,classifier,Knearestneighbor(
KNN).
Type : Research paper
Published : Volume-11,Issue-3
DOIONLINE NO - IJASEAT-IRAJ-DOIONLINE-20262
View Here
Copyright: © Institute of Research and Journals
|
|
| |
|
PDF |
| |
Viewed - 20 |
| |
Published on 2023-12-18 |
|