(IJCSIS) International Journal of Computer Science and Information Security, Vol. 13, No. 9, September 2015 LDA-PAFF: Linear Discriminate Analysis Based Personal Authentication using Finger Vein and Face Images Manjunathswamy B E 1 , Dr Thriveni J 1 , Dr Venugopal K R 1 1 Department of Computer Science and Engineering, University Visvesvaraya College of Engineering, Bangalore, India Abstract— Biometric based identifications are widely used for individuals personnel identification in recognition system. The unimodal recognition systems currently suffer from noisy data, spoofing attacks, biometric sensor data quality and many more. Robust personnel recognition can be achieved considering multimodal biometric traits. In this paper the LDA(Linear Discriminate analysis) based Personnel Authentication using Finger vein and Face Images (LDA-PAFF) is introduced considering the Finger Vein and Face biometric traits. The Magnitude and Phase features obtained from Gabor Kernels is considered to define the biometric traits of personnel. The biometric feature space is reduced using Fischer Score and Linear Discriminate Analysis. Personnel recognition is achieved using the weighted K-nearest neighbor classifier. The experimental study presented in the paper considers the (Group of Machine Learning and Applications, Shandong University-Homologous Multimodal Traits) SDUMLA-HMT multimodal biometric dataset. The performance of the LDA-PAFF is compared with the existing recognition systems and the performance improvement is proved through the results obtained. Keywords-SDUMLA_HMT; LDA-PAFF; Phase; Magnitude; fisher Score I. INTRODUCTION The use of biometrics to identify personnel is widely adopted in our day-to-day scenario. A biometric recognition system identifies an each personnel using one or more specific physiological characteristics possessed by the individuals [1]. If one physiological characteristics is considered for recognition then they are termed as unimodal recognition systems. When multiple or a combination of personnel biometrics are considered then they are termed as multimodal biometric recognition systems. Enrollment and verification of authorized personnel are the important functions of the recognition systems. The recognition systems enroll authorized personnel based on the data provided from the biometric sensors and store the data for future verification or matching. During verification the recognition systems validates with the existing whether the biometric data presented is valid or invalid. Predominantly unimodal systems are adopted for personnel identification [2]. Key challenges in unimodal biometic systems: The unimodal biometric recognition systems currently used in day-to-day activities suffers from large number of drawbacks [2][3][4]. Biometric recognition systems solely rely on the data provided in the biometric sensors. The data input provided to the recognition systems from the sensors are generally noisy in nature which can affect the verification results and also cause faulty enrollment techniques. The illumination variation for face recognition systems is one such example. Interpersonal biometric similarities is another drawback of unimodal biometric systems [4]. Unimodal biometric recognition system presented in the research work using the finger print [5] it clearly illustrates the biometric similarity problem. Spoofing attacks are the common causes in unimodal recognition systems. Spoofing attacks are commonly noticed when biometrics like signature, voice, face and finger prints are considered [2] in the recognition system. Motivation: Multimodal biometric recognition systems is used to overcome the drawbacks of the unimodal recognition systems and have proved to be 142 https://sites.google.com/site/ijcsis/ ISSN 1947-5500