This document discusses biometric authentication systems and proposes two new models to improve security. The models integrate biometric authentication with cryptographic techniques like RSA and MD5 to better detect "spoof attacks" where an imposter tries to access the system by spoofing a genuine user's biometric data. The first model uses RSA to generate secret keys from user-provided random numbers. The second model authenticates users by having them answer questions from a predefined set. Together these models aim to provide stronger authentication by verifying a user is alive and not an imposter trying to spoof the system.
This document discusses biometric authentication systems and proposes two new models to improve security. The models integrate biometric authentication with cryptographic techniques like RSA and MD5 to better detect "spoof attacks" where an imposter tries to access the system by spoofing a genuine user's biometric data. The first model uses RSA to generate secret keys from user-provided random numbers. The second model authenticates users by having them answer questions from a predefined set. Together these models aim to provide stronger authentication by verifying a user is alive and not an imposter trying to spoof the system.
This document discusses biometric authentication systems and proposes two new models to improve security. The models integrate biometric authentication with cryptographic techniques like RSA and MD5 to better detect "spoof attacks" where an imposter tries to access the system by spoofing a genuine user's biometric data. The first model uses RSA to generate secret keys from user-provided random numbers. The second model authenticates users by having them answer questions from a predefined set. Together these models aim to provide stronger authentication by verifying a user is alive and not an imposter trying to spoof the system.
This document discusses biometric authentication systems and proposes two new models to improve security. The models integrate biometric authentication with cryptographic techniques like RSA and MD5 to better detect "spoof attacks" where an imposter tries to access the system by spoofing a genuine user's biometric data. The first model uses RSA to generate secret keys from user-provided random numbers. The second model authenticates users by having them answer questions from a predefined set. Together these models aim to provide stronger authentication by verifying a user is alive and not an imposter trying to spoof the system.
Copyright:
Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online from Scribd
Download as pdf or txt
You are on page 1of 6
$EVWUDFW Biometric technologies` have become the basis for
highly secure identification and personnel verification in
authentication and identification systems, based on the physical identifiable details or behavioral characteristics which cannot be imposter easily. Hence, the system has been treated to be the tough security measures to identify the user being genuine by avoiding the third party attack. But these systems are unable to judge whether the user who is providing the required authentication details is an impostor or genuine. To eliminate this flaw in biometric based security systems, the model proposed in this paper can be embedded in existing security systems to prevent the spoof attack. The proposed model comprises of two independent models namely Cryptographic based authentication system (model 1) and Text prompted authentication system (model 2). Model 1 has been designed by the most efficient cryptographic algorithm RSA which generates a secret key by using a pseudo-random number given by user for secure system access. In model 2, the user is made interactive with the system by prompting a question from a set of 15 questions, and authenticating the user with the help of correct answer provided by user. To provide higher level of security, MD5 (Message Digest 5) algorithm is integrated with RSA. The developed models provide promising level of security along with a biometric trait. Key words: 8VHU $XWKHQWLFDWLRQ%LRPHWULFWUDLWV$OLYHQHVV 'HWHFWLRQAntiSpoofing&LSKHUWH[W6HFXUHGGDWD 1. INTRODUCTION In the networking society, the needs oI the user are catered with the help oI secure inIormation/data transIer. But the problem that rises around this is whether the set oI inIormation/data is used by right persons? in right way? So this is answered in every security domain, by providing inIormation/data to the authenticated user through security measures. The user authentication is a challenging task in any security system. One oI the method to authenticate the user with the help oI biometric evidence. Biometric system are automated methods oI veriIying or recognizing the identity oI a person based on the physiological characteristics, like a Iingerprint, Iace pattern etc. and/or some aspects oI behavior, like handwriting, speech patterns etc. User authentication can be made with assumption that no two individuals` biometric characteristics may seem to be identical. In voice based systems, vocal tract shapes, larynx sizes, and other parts oI their voice production organs are diIIerent. Every user can also produce characteristic like manner oI speaking, use oI a particular accent, rhythm, intonation style, pronunciation patterns, choice oI vocabulary, etc. with voice oI an user In retinal based system, the nervous system connected within the eye may be diIIerent. In Iingerprint, the ridges in the Iingers may vary Irom one to another. In order to provide higher level oI security, the biometric systems are to be integrated with additional security measures. Most oI the biometric systems are attacked by the imposter with spooI` attack. The spooI attack Iools the biometric system with identity theIt and other related criminal activities and also gain access to soItware or others personal inIormation. The phenomenon spooI is a counterIeit biometric which is used in an attempt to circumvent a biometric sensor. DiIIerentiating a genuine biometric trait presented Irom a live person versus some other means oI same type is called spooI detection. Hence, the act oI sensitivity 'aliveness signs, such as detection oI heart pulses. In some areas oI research, the term aliveness detection is synonymous with spooI detection. In other areas oI research, aliveness detection is the more limited problem oI distinguishing a live human trait Irom a non-living human trait and in some systems aliveness detection is deIined as the sensing oI vitality signs. SpooI detection can occur beIore biometric data is collected or during data processing. Whenever a new data is entered, in a decoupled system no biometric data is collected until the spooI detection method is satisIied that a live human is present. For identiIying the aliveness detection and Iurther to authenticate user to achieve higher level oI security, we explore the idea oI biometric system which can be deployed in all public interactive and sensitive environments Ior strengthening the security system. The proposed model is Iramed to integrate the biometric authentication, and two independent models designed using the Iastest MD5 and the most secured RSA techniques which gives the betterment oI the existing biometric system in advancement oI providing security by identiIying the user being genuine or imposter. Further, in section 2 describes the methods and techniques available Ior user authentication using diIIerent biometric traits and methods Ior identiIying the genuine user. Section 3 discusses the proposed security model section 4 deIines the algorithm and Iinally section 5 presents the results oI the proposed work. 2. LITERATURE REVIEW Most oI the researchers addressed the issues in biometric systems as a security measure by considering many scenarios where the user`s biometric details oI Iinger/hand/palm/eye/voice (input) etc. is perceived and is acquired by an acquisition devices to represent them into digital signals. The research works carried out on various biometric traits are summarized in the Iollowing; in our previous work, the methodology proposed in |1| was designed Ior A Cryptographic Approach to Prevent a SpooI Attack Ior Secure InIormation Retrieval in a Biometric System Ravi Hosur Department oI CSE Basaweshwar Engineering College Bagalkot - 587102, Karnataka, India hosuravigmail.com Sanjeevakumar M. Hatture Department oI CSE Basaweshwar Engineering College Bagalkot - 587102, Karnataka, India smhattureyahoo.com Rashmi P. Karchi Department oI MCA Basaweshwar Engineering College Bagalkot - 587102, Karnataka, India rashmikarchigmail.com 326 978-1-4673-4805-8/12/$31.00 c 2012 IEEE the security system by using two stage authentication processes. The model works Ior the wide area networks like public sectors. The issue oI authenticating the genuine user (aliveness) when third party is using the system instead oI authenticated user is addressed in this paper. Further, our model in |2| was designed Ior LAN security system by using text-dependent speaker identiIication system. The system was modeled only to work in limited area like intranet or LAN system. In this paper we extend the system Ior WAN system by considering the issues related to security measures concerned to internet also. The method |3| works with speech and Electrocardiogram (ECG) signals using wavelet transIorm Ior cryptographic key generation based on the uniqueness and quasi-stationary behavior oI ECG and speech signals Ior an individual. The designed system is simulated to increase the perIormance with a report oI False Acceptance Rate (FAR) oI 1.27 and a False Rejection Rate (FRR) oI 10.62 Ior the system. The paper does not compromise with the keys oI a group or a corporation that could happen in the case oI maintaining a centralized database with the biometric inIormation oI all users but can compromise with the key oI user hacked by the third party. The model |4| detects the spooI Iingerprint attacks in the biometric system is on the observation that, real and spooI Iingerprints exhibit diIIerent textural characters based on structural, orientation roughness, smoothness and regularity diIIerences oI diverse regions in a Iingerprint image where the images are captured by Local Binary Pattern(LBP) histograms. The dimensions are integrated as Ieature set which is reduced by running Pudil`s Sequential Forward Floating Selection (SFFS) algorithm. A new image (single) based method utilizing integrated gray level texture and wavelet energy inIormation Ior spooI Iinger detection is presented. In the methodology |5|, wireless Local Area Networks(WLANs) that use MAC Iiltering to allow stations with registered MAC addresses to use the network where the attacker has many tools like AirJack, wireshark to capture the packets in WLANs and Iind authorized MAC address. The attacker masquerades as an authorized station and can launch denial oI service attack. The work presented a power hopping technique which can be used by Access Point (AP) to discern the authorized packets Irom the masquerading packets and thus deny the attacker Irom using the system. The AP learns about the signal strength about various locations and also AP learns about the noise to signal ratio. The technique could more consumption oI energy because even iI the station is near AP, in order to meet expected power level, station might use more power than the optimal power needed. The method |6| extracts the Iingerprint Ieature oI cognitive user to get the Received Signal Strength (RSS). The extracted Eigen values can be used as the input vectors oI the Back Propagation (BP) Neural Network which are trained by the honest users` RSS. Simulation results shows that the scheme can eIIectively detect identity spooI attack with a low Ialse alarm rate and miss alarm rate. The process extracts Ieature inIormation RSS value, and classes these data through Back Propagation neural network. The work |7| exhibits a worst case scenario that the attacker is able to Iabricate an exact replica oI genuine biometric trait which was simulated by assuming that the matching scores distribution oI Iake traits is identical to the one oI genuine users. So the biomedical system used does make a Iace and Iingerprint matcher whose scores are Iused using the well-known sum, product, weighted sum and Like-Li-hood Ration (LLR) rules and, investigates whether a realistic spooI attack against both modalities can allow the attacker to crack multi-modal system. Results in robust and perIormance Iactors under a spooIing attack against all traits are still unacceptable Ior security applications. In other words, they can be cracked by spooIing all the Iused traits even when he attacker is not able to Iabricate an exact replica oI the genuine user`s traits. In the method proposed |8|, here an oIIline signature recognition system is built along with speaker recognition system Ior Ieature extraction oI the data perceived Irom the user with the countable data. The sum rule is used Ior the Iusion oI biometric scores and experimentally resulted in the eIIicacy oI multimodal biometric system using speech and signature Ieatures when the data is aIIected by noise. With the results oI the method implemented |9| justiIies that the RSA algorithm is stronger than DiIIie-Hellman method with the key length parameter and is more securitized during timing attacks, proves to be the most highly rated algorithm regarding the encryption ratio, speed and tunability parameter also. These all suggest that the RSA algorithm is tunable and Ilexible to diIIerent type oI applications and requirements and best suitable Ior highly valuable multimedia applications. The scheme in |10| presented some oI the cryptographic key generation schemes so that the biometric system with traits like Iingerprint, iris, retina, etc., are used in rendering security to the inIormation or key able to satisIy the growing demand oI the security in the current real world applications and the growing demand against the security. The paper |11| presents a hand geometry based user identiIication using spectral properties. The hand images used are collected by a GPDS150 hand database. The user identiIication is perIormed using back propagation multilayer perceptron network (BPNN). The correct identiIication rate about 98 is achieved. The technique |12| is revolved over creating an elliptical curve points Irom Iinger print oI the user, where the data are collected in the Iorm oI minutiae, and using the minutiae points the method generates an elliptical curve with the help oI elliptic curve cryptography generation algorithm. This elliptic curve generated helps in validating the user Ior authentication purpose. In this technique, Iinger`s minutiae is converted into an elliptical curve using elliptic curve algorithm where the extraction is done by binarization, thinning, noise removal and minutiae extraction, and elliptic curve points generation steps with deIined minutiae templates like termination, biIurcation, lake, independent ridge, point or island, spur and crossover. The author states that the proposal can be used in document authentication like passport and driving license where the biometrics can be used. 2012 World Congress on Information and Communication Technologies 327 The Iinger vein location method |13| takes the vein as an input data that treats the data as a high counterIeiting capability in comparison to other biometric trait used Ior identiIication. The method also designs a Iinger vein imaging device Ior establishing templates oI images oI Iinger veins oI diIIerent users with diIIerent patterns to conduct the personal identiIication, and the database is created with all these data. The extraction oI patterns is done by using the steps oI image processing and then compared with the database values Ior validating the user Ior authentication. One oI the biometric traits Ior authenticating a user is Iace |14|; used Ior veriIication oI user Ior authentication by using eIIicient Ieature extraction and an error minimizing template transIormations. The Ieature has been generated as a template is the result oI Gradient Descent Algorithm based on error rate Iormulation. The algorithms like Multiple Random Projection (MRP) Ior Ieature level extraction, Total Error Rate (TER) Ior veriIication are used with AR Iace and BERC visual Iace databases in the process oI proposed methodology. The multimodal technique |15| is built with integration oI biometric traits like iris and palm print based on Wavelet Packet Analysis. The Ieatures considered eye is its texture and principle lines, wrinkles and ridges on the palm that are encoded into a compact sequence oI 2-D wavelet packet coeIIicients that generates a Ieature vector code. Then by using Wavelet Packet TransIorm texture and recognition oI traits considered are analyzed and identiIication by Hamming distance. The proposal |16| bases on the Ieatures oI singular creasure, puts Iorward multi-biometric identiIication technology based on Iinger shape and palm print. The method extracts the Ieatures including length and average width oI Iour Iingers and with palm Ieatures oI palm rectangular region, where the validation is done using hierarchic mode with coarse and reIined matching methods by actively increasing the Iaster match. From the above cited works on cryptography, biometric identiIication methods, etc. using diIIerent schemes provide security to the data Ior storage but Iail to detect the user who is providing the identiIication details are imposter or genuine. Hence, there is a scope to address an issue on the aliveness detection and authenticate user in security systems whether the details provided by the user is being genuine or not by the live` users. Resulting in more security to the system and save the system Irom the third party attack. 3. PROPOSEDSYSTEM To provide higher level oI security, it is planned to integrate the techniques oI biometric security and cryptography. Further, the proposed system also veriIy Ior the genuine user by prompting the conIidential random question to the user. The proposed system tries to block the imposter at diIIerent stages. The details oI the work carried towards this direction is presented below; The proposed model is Iramed with two models that depend on the type oI user being identiIied (aliveness/genuineness), by integrating the proposed model between the authenticating system and the database so that the user is judged to be genuine. The proposed model is shown in Iigure 3.1 Figure 3.1: Enhanced Antispoofing Model Working Principle The user is authenticated with his/her biometric trait like iris, Iinger print, palm print, etc. given as details to the system in the basic authentication step. The system will permit the user to enter the system with maximum three attempts where each time the user has to provide diIIerent biometric trait randomly (without repetition). II the authentication Iails with all three attempts, then the system assumes that the user is a counterIeit. Otherwise the user is asked to provide a NONCE (pseudo random number oI 4 digits) which in turn the system asks Ior a secret code or an identiIication number. II the number generated due to secret code and NONCE already exists in the CACHE (repeated) then it`s a counterIeit data then model 1 is activated else the model 2 is activated to provide Iurther security. In basic authentication step all details oI the user are being stored in database and encrypted by using MD5 algorithm (known to be the best cryptographic algorithm). These cryptographic parts are very hard to break because the data will be stored is not the real as provided by user during his registration to the system but they are in other Iorm called CIPHERs. MD5 algorithm uses one oI the details (attribute) as its key Ior generating CIPHER. Model - 1: Cryptographic based authentication system The resultant oI mismatch with the key (i.e. resultant oI SECRET Key and random number) makes this module activate. When the user wishes to access data, then he/she needs to enter a SECRET code (4-digits key a pseudo-random number), which will be encrypted and sent immediately to the user`s registered mobile number or email account similar to one-time password. Then the system asks the user to re-enter the SECRET code (encrypted) to the system. II the entered code by user is valid then access to the 328 2012 World Congress on Information and Communication Technologies system will be provided otherwise user will be given maximum oI three attempts to prove that he is the real user (genuine). Figure 3.2: Cryptographic based authentication system Module - 2: Text prompted authentication system This module is activated (i.e. Iigure 3.3) iI there is no repetition oI the value at the basic authentication step. In this module, when user selects a transaction to perIorm on the data then user needs to answer a random question generated by the system (among the 15 questions where these answers are registered by the user initially). Once the user provided answer is correct then 8-digit alphanumeric value will be sent to user`s mobile/email. When the user enters 8-digit alphanumeric value, is valid then user can access and perIorm the actions on the data. Otherwise the system will generate diIIerent question oI maximum oI three. The user need to prove that he is the real user (genuine). In this module every question occurs only once and never repeats until all questions complete their turn oI occurrence. Once these all questions complete their turn (all 15) the system generate diIIerent set oI 15 questions, Ior them user need to provide the answers, which are registered against the same user`s data with encryption. Figure 3.3: Text prompted authentication system 4. Algorithm for the proposed model 4.4.1 Algorithm Proposedsystem () Step 1: Registration oI the user to security system using Module Register_User () Step 2: AIter registration the system service can be used by logging in with Log_In () Step 3: Stop 4.4.2 Algorithm Register_User () Step 1: User needs to provide all biometric traits like Iris, Iinger print, palm print, etc. and his required details like Iirst name, last name, Designation, address, etc. Step 2: Once the details are provided and submitted, they are stored to the database Step 3: User account will be created with his username and password given at the time oI registration Step 4: Stop 4.4.3 Algorithm Log_In () Step 1: when the system validates the user being genuine by identiIying his biometric trait given then second authentication stage with his/her valid username and password, he/she will be permitted to use the system Step 2: AIter logging in user can perIorm diIIerent transactions depending on his choice by giving suitable required details to the system like answers to the registered questions Step 3: II the answer given is correct then it gets access to the system`s transaction page, otherwise it goes to Cipher text generation page (model-2 in proposed model) and asks to enter a random number in turn the system regenerates the cipher and sends it to user`s mobile/email where the user has to enter within stipulated time Step 4: iI cipher is entered in time with correct data then gets access to the system`s transaction page where the user really intended to go. Step 5: Stop 5. Experimental Results The proposed model is tested Ior accuracy with 100 diIIerent users using the Iingerprint biometric traits and accessing the designed system to IulIill their requirements. The biometric system authenticates the genuine user with Iingerprint trait and Iurther activates the either the model 1 or model 2. MD5 algorithm at the user side encrypts the details in to other Iorm (cipher) by using one oI the details (attributes provided by user) as encryption key Ior conversion during storage, resists Irom an imposter. During validation the user details, the algorithm decrypts the stored data and compares with user data Ior authentication. The role oI RSA algorithm is hiding the process details by encryption and produces the required results oI the user using decryption method i.e., to know whether the genuine user is 2012 World Congress on Information and Communication Technologies 329 perIorming a transaction the user is interactively made to participate with the system during the process oI identiIication and authentication. The Iollowing are the results that produced during the process to Iind the accuracy in regard with the counterIeit details; Table 1. Accuracy of the proposed system Application Parameters Number of Users Accuracy in of detection counterfeit Authentication using MD5 100 93.45 Biometric trait 100 94.65 Encryption/Decryption Using RSA 100 96.25 From the Table 1, the experimentation has been carried Ior 100 diIIerent user sets. In the results, Iirst parameter tested is the basic authentication mode said to be the most rigid step, where the counterIeit user Iails to login to the system because he/she Iails to decrypt the details entered by the user i.e., his username and password. Hence result is more than 90 because oI the eIIiciency oI MD5 algorithm. Then the second parameter results around 94 because there can be a chance oI moldings oI user biometric trait that may happen due to impostor creation or due to the virtual storage oI data in the system buIIers during power Iailures even though the biometric system has been developed with interesting techniques. The transaction page that produces a questionnaire Ior the user depending on the randomly selected question to be answered by the user using pseudo-random number generator never gives the intruder an idea by encrypting the text using RSA algorithm. The code sent to the user will be in the Iorm oI encryption aIter entered by the user and will be decrypted during validation. The algorithm is proved superIicially strong in encryption and does- not allow breaking the system and guess the system secrete. Hence the proposed application is expected to produce averaging about more than 90 results compared to existing biometric security systems. 6. CONCLUSION The proposed system integrates the concepts oI cryptographic and biometric to provide higher level oI security. The developed model can be eIIectively deployed in existing biometric security system to prevent the spooI attack by providing the aliveness detection and improve the protection oI data access Irom intruders. The proposed system has been implemented by integrating standard algorithm that are enriched with materialistic approach along with parameters like speed, accuracy, eIIiciency, tunability, etc. this integration results to produce a more strategic security system that asks a question whenever it is in an ambiguous state. In Iuture the system improves the biometric system capability by using multiple evidences oI the user and advanced cryptographic techniques to cater the demand Ior higher level oI security to the system. REFERENCES 1]. Ravi Hosur, Sanjeevakumar M Hatture and Rashmi P Karchi, ' AntispooIing Model Ior Secure InIormation Retrieval in a Network Society, proceedings oI International ConIerence on InIormation and Network Technology (ICINT 2012), 2012, pp 43-48 2]. Ravi Hosur, Sanjeevakumar M Hatture and Rashmi P Karchi, ' A Model to Prevent a SpooI Attack using Text- dependent Speaker IdentiIication, proceedings oI Advances in Computer Science and InIormation Security(ACSIS`12), 2012, pp 73-77 3]. H. A. Garcia-Baleon V. Alarcon-Aquino O. Starostenko J. F. Ramirez-Cruz, 'Biomodal Biometric System Ior cryptographic Key Generation Using Wavelet TransIorm, Proceeding ENC '09 Proceedings oI the 2009 Mexican International ConIerence on Computer Science IEEE Computer Society Washington, DC, pp 185-196, 2009 4]. Shankar Bhausaheb Nikam Suneeta Agarwal, 'Texture and Wavelet-Based SpooI Fingerprint Detection Ior Fingerprint Biometric Systems, First International ConIerence on Emerging Trends in Engineering and Technology, icetet, pp.675-680 2008 5]. Vijayakrishnan Nagarajan, Vetri Arasan, Dijiang Huang, 'Using Power Hopping to Counter MAC SpooI Attacks in WLAN, Proceeding CCNC'10 Proceedings oI the 7th IEEE conIerence on Consumer communications and networking conIerence IEEE Press Piscataway, NJ, USA 2010 6]. Shan Kang, Naiwen Chen, Mi Yan, Xiaoxiao Chen, 'Detecting Identity-SpooI Attack Based on BP Network in Cognitive Radio Network, pp 1603 1606, 2011 7]. Zahid Akhtar, Battista Biggio, Giorgio Fumera, and Gian Luca Marcialis, 'Robustness oI Multi-modal Biometric Systems under Realistic SpooI Attacks against All Traits, IEEE Workshop on Biometric Measurements and Systems Ior Security and Medical Applications (BioMS 2011), Milan, Italy, pp. 5-10, 2011 8]. Kartik, P.; Mahadeva Prasanna, S.R.; Vara Prasad, R., 'Multimodal Biometric Person Authentication System using Speech and Signature Features, TENCON 2008 - 2008 IEEE Region 10 ConIerence pp 1 6, 2009 9]. Al. Jeeva, Dr. V. P. Palanisamy and K. Kanagaram, 'Comparative Analysis oI PerIormance EIIiciency and Security Measurements oI some Encryption Algorithms, International Journal oI Engineering Research and Applications (IJERA) ISSN: 2248-9622, Vol. 2, Issue 3, May-June 2012, pp. 3033-3037 10]. Mr. P Balakumar, Dr. R Venkatesan, 'A Survey on Biometrics based Cryptographic Key Generation Schemes, IRACST - International Journal oI Computer Science and InIormation Technology & Security 330 2012 World Congress on Information and Communication Technologies (IJCSITS), ISSN: 2249-9555, Vol. 2, No. 1, 2012, pp. 80-85 11]. Shanmukhappa A. Angadi, Sanjeevakumar M. Hatture, 'A Novel Spectral Graph Theoretic Approach To User IdentiIication Using Hand Geometry, International Journal oI Machine Intelligence (IJMI), Volume 3, Issue 4, December 2011, pp. 282-288. 12]. B. Thiruvaimalar Nathan, Dr. R. Meenakumari and S. Usha, 'Formation oI Elliptic Curve Using Finger Print Ior Network Security, International ConIerence on Process Automation, Control and Computing (PACC), 2011, pp 1-5 13]. Wenming Yang, Qing Rao and Qingmin Liao, 'Personal IdentiIication Ior Single Sample Using Finger Vein Location and Direction Coding, International ConIerence on Hand-Based Biometrics (ICHB), 2011, pp 1-6 14]. Youngsung Kim and Kar-Ann Toh, 'A Method to Enhance Face Biometric Security, First IEEE International ConIerence on Biometrics: Theory, Applications, and Systems, BTAS 2007. 15]. Hariprasath S. and Prabhakar T.N., 'Multimodal Biometric Recognition Using Iris Feature Extraction and Palmprint Features, IEEE International ConIerence on Advances in Engineering, Science and Management (ICAESM), 2012, pp 174-179 16]. Lang Zhai and Qi Hu, 'The Research oI Double- biometric IdentiIication Technology Based on Finger Geometry and Palm print, 2nd International ConIerence on ArtiIicial Intelligence, Management Science and Electronic Commerce (AIMSEC), 2011, pp 3530-3533 Authors: 1]. Mr. Ravi Hosur Completed BE Irom Visvesvaraya Technological University Belgaum. Currently pursuing Master oI Technology (MTech) in Basveshwar Engineering College Bagalkot, extension center oI Visvesvaraya Technological University (VTU) Belgaum. He is the student member oI IEEE, liIe member oI Computer Society oI India (CSI) and Indian Society Ior Technical Education (ISTE). His areas oI interests are InIormation security, Data Mining, Algorithm design and Analysis, Database security. 2]. Prof. Sanjeevakumar M. Hatture Completed BE Irom Karnataka University Dharwad, Master oI Technology (MTech) Irom Visvesvaraya Technological University Belgaum and currently pursuing doctoral at Basaveshwar Engineering College Bagalkot research center under Visvesvaraya Technological University Belgaum. He is the liIe member oI Institute oI Engineers (IEI) and Indian Society Ior Technical Education (ISTE). His areas oI interests are Image processing, Pattern Recognition, Biometric security and Network Security. 3]. Prof. Rashmi P. Karchi Completed MSc (Stat) Irom Karnataka University Dharwad and Master oI Technology (MTech) Irom University oI Mysore. She is the liIe member oI Institute oI Engineers (IEI) and Indian Society Ior Technical Education (ISTE). Her areas oI interests are Image processing, pattern recognition. 2012 World Congress on Information and Communication Technologies 331