K-MEANS BASED MULTIMODAL BIOMETRIC AUTHENTICATION USING FINGERPRINT AND FINGER KNUCKLE PRINT WITH FEATURE LEVEL FUSION

In general, identification and verification are done by passwords, pin number, etc.,
which are easily cracked by others. To overcome this issue, biometrics has been introduced as a
unique tool to authenticate an individual person. Biometric is a quantity which consists of
individual physical characteristics that provide more authentication and security than the
password, pin number, etc. Nevertheless, unimodal biometric suffers from noise, intra class
variations, spoof attacks, non-universality and some other attacks. In order to avoid these attacks,
the multimodal biometrics, i.e. a combination of more modalities is adapted. Hence this paper has
focused on the integration of fingerprint and Finger Knuckle Print (FKP) with feature level fusion.
The features of Fingerprint and (FKP) are extracted. The feature values of fingerprint using
Discrete Wavelet Transform and the key points of FKP are clustered using K-Means clustering
algorithm and their values are fused. The fused values of K-Means clustering algorithm is stored in
a database which is compared with the query fingerprint and FKP K-Means centroid fused values
to prove the recognition and authentication. The comparison is based on the XOR operation.
Hence this paper provides a multimodal biometric recognition method to provide authentication
with feature level fusion. Results are performed on the PolyU FKP database and FVC 2004
fingerprint database to check the Genuine Acceptance Rate (GAR) of the proposed multimodal
biometric recognition method. The proposed multimodal biometric system provides authentication
and security using K-Means clustering algorithm with GAR=99.4%, FRR=0.6% and FAR=0%
with security of 128 bits for each modality.