BIOMETRICS SS On Ta ; 4 Authentication is a fundamental component op teeing interaction with the Computers ~* Traditional means & authentication primarily pass and pas nal Identification Numbers (Pin) have until recently, «deminated,.completly, and are likely to. vyemain escentially for wears to come» However, | aml Axtthentication technologies, copable of, Prov iP higher degree of Security % Biornetrics Ave one Such ere authentication technology - Benefits of Biometrics verses “Mraditional, Authentication] methods: i 4g Passwords, PINS have a number of problems that call into question their Suitability for modern applications, porticolarly high Security applications such as access to online financial transactonc ey medical data A-Increased Secunity: Biometrwics Can provide a Greater deqree of secuvity than traditional authentication methods i-e» vesovyces ave only accessible to authorized users and ave kept protected From. unauthorized. users Biometrics data cannot be guessed oY stored in the same fashion as @ password or PIN. ? oh ) i Buy a. Tneveased Convenience: One: .0f. the reasons passwords cre Kept. simple is that they ave easly Fovgolten.Biometrics ave difficult AF not impossible to forget, eo con Offey much Greater Convenience Systems . based_on ae t) moltiple password Highly sensitive information can move ‘readily be _ Made available on a biometrical ‘protected network than on O one protected by passwords can’ increase’ sey and enteyprise Convenience, Qé users lcd access’ dthewwise "Protected | information 7 withovt the need of human Intervention 3.Increased Accountability: : Osing biometrics to secuve Cornputess and facilities eliminates: phenomena such ac buddy / spugchin And provides a high deqvee of Sertalnityg as to what usev Accessed what Computer at whi time .Even if the Cee and ~ ting capabilities cf 'a Suctem ave warvely wsed, the fact’ that they exist’ of ten ervey as Ken “effective dettewent Benefits ts lof Biometrics ’ ‘ip V lenea dea, Sjstern.” “In Tdentification systems, Biometrics Can be bs fov Security, Convenience And accountability - Espec when they are deployed to a ‘modest. Humbey of OSENS: However Identification Systems AVE more" ‘oFle deployed in large Seale ebvivdmmentro oct a Ae 4 Froud Detection: t Identification edd 0 ave dpi ed to deter whether a persone bidmetale! information exis? tore than once in @ database - By locatin ong tolentifying Individuals who toKe hive ‘already! Veqisteved for a teats oY Service, biometvicg can veduce fYaud+ Fe siehgy2 Froud, Dekenonce: Move than fraud detection, Fraud oleterrance Is & primary benefit tn lange Scale en fieot Ona Bs “The presence of biometric identification technolog: can detev individual Exam attem pting to. envotl mot fipte times in a public benefit systems, then the peal agency has saved money and ensured the Inteqrit of vecovde -In the absence of btometyics, theve is no effecient woy of identi fying duplicate applicants ox ssegistvants and rt is therefore difficult to detey such. Opplicotions- Biometrics pIODE LHI CS Biometrics ts the automated ose of physiological ov behavioural chorvacterstics tO determine ov verity identi€ Automated Use: © * Brometyic Lechnotogies wre automated - computers 08 machines are used 'to verity oy iden tity ov determine | _ ' | identity Papuge behavioural ow Physiclog: cal charaict- | ~exstics AS the process 1s Qutomated , biome tric authentication nevaally vequites enly a few seconde and biometric J, doth “ Systeme ave able to comme thousand, Of vecords pe Secor -A sustens where in a Sey laces pesiroing Physiolog cal oy rou OF Behavouval Charracterstics. Biometrics are based on the mea physiological and behavioural characterstics- Fingey sean, factal scan, hand'séan, Retina ean oe . considered as physic Worement OF alisting tive eal biometrics, based On” dlivect easuremest MOF a Poot_OF the human beelyNoice scan and Signatuve scan are Considered ac behavioural ‘biornetrice , they ane based on measurel and data’ veceived devived from an action ‘and therefore indirectly measired ehanacterstics of he ‘beay > Verification and Ide: Username user Presents est dag | olloitt-- bi ta data My nome i testias Test 4 Jotrot.--4 Does my biometric match the date Stoved fox that username 2 Fam presentini Seon leet cared What name ic astostated with the biometric data aN The user's biometric clata is compared against muiti ple oseys biometric data 1 x Verification systems answer the question 7 “Am F who J Claimed to be 2”, by vequited that a User claim an identity inoyder fer.a, biometric compavigon to be performed- Aftex a user claims a sera uses provides biometvic data tohich ic then Cee against bic emolled biometic dakNice scan and Signatuve scan @re Considered ar behavioural “biometrics , they axe based on measure and data’ veceived devived from’ an action ‘and thevelore indivectly measured ehanacterstics of hy ‘beay > Verification and Ide: user presents Test aa —biometiia data, My name ic testiaz Test aq Does my biometric Match the date Stoved fow that useTName 3 Fam presentini a biometric data that name “ic astoctated with the biometric data aan The user's biometric clata is compared against multiple osevs biometric data 1 x Verification systems answey the question P "am T coho I” Claimeel to be 2”, by yequited that 4 a User claim an identity inoyder fer.a, biometric comparigon to be performed- Aftex a user claims @ Sera uset provides biometvic data tohich i¢ then Qernponed agpusst hic emolled biometric da* Depending on the type of biometric systems the helen tate that a wsey claims might be @ window's Username , @ given mame oF an id number, the answey returned by the system is Match 0¥ NOMatch verification systems Can contatn millions of biometric Wecords, but are alwayl predicted on a user's biometric. data being featchecdt against only his cen emotled biometiic date » Verification i¢ always yelerved as ht [one bo-one] x The precess of Providing a user name and Biometric data s iS referred tO 2S AN authentication ¥ Identification systems answer the question “Who aml? | and donot vequived that a User, cClagmsan identity before biometric ‘comparison take place * The ucey Provides bic biometric data which is. com to data from a “number of Users ineides to find a match The answer returned by the Template eee - bh ee the Biometyic oa Peete ee jencovs cera shold ae algoithen ae — Wenders resubmit, data x A vsex initially envolls in biometric systems by oe biometri¢ data , whith is converted into a template x Templates ae stored In biometric systems fox the purpose of subsequent copay a Trovdey to be verified ox identified after a ! emrollment , the User provides biometyic data, whig is conveyted twnto a template | x The verification template is componed With one | oy move EMeliment templatesx The vesult of a, Componison betuoeen biometric templates 1¢ endeved as a score oy confidence level, whieh i¢ compared to a ttreshold Used fer a Specific technology s systems USeY Ov transaction % If the ecove exceeds the threshold the Comparison is a match and that ‘yesult 1 Evansmitted * IF the ‘Stove does not meet the thyeshold, the comparisen| ts MOE a match and that Yesult is transmitted Emoliment end Template Creation “the process by which a users biometric data i¢ mitiatl pited, Accessed, processed and Stored * in the form OF a template for engoing use in a Biometyic System, 1¢ called enveliment i ¥ Quolity emeliment 15 a critical factcy in the lon -teym accuracy of biometric systems dow eon : enrollment m lead t> high evry vyatec, ercl Ly falce match rete ‘and false” Nem match rate Presentation After a User provides uohat evev’ peysonat information ts required to begin’ emoliment, such ac name or userid, Ose presents biometyic data Biometric Data: “The biometvic data provided by the Uses, is an G@nprockssed, image cy Yecoyding Of a Chavacterstic tis alse catled as yaw Ridmetric data © Lite the Biometric system, O Lsex may. 4 , necd to present biometric data Several timec Inovdey to’ envoll « Enyollment tegqones. the Creation of an identifier such as Usey Name ov. id,Biometyic Technology Data type Fingey scan Fingerprint image Voice’ Scan. Voice Tecordin, Face scan « €aciat im Tris stan” Tvs image “Retina scan Retina Ima: “Hand scan BO tmage of the 5 Cop and sides) " SENG scan Im oge of Signa te Features “Extraction: =) + * “The automated process of locating and encedi distinctive chayacterstics fYom biometric data and inovdey bo generate a bemptate is called | feature, extvaction In voice ¢can technoliagy + the process includes filter certain frequencies and patterns Ia fin ey Scan technology process incwdes 4 oc of the vidges to the width of a Single pixel Templates a termplate is Q Smalt file devived from the disti features Of a Users ‘biometric dalayused to pevform biometric matches ‘ ~| x Biometyic Systems Stove and Compane biome ti compares Not biometric data x Most templates occupy 4 bytes to vkilo byte x There ts no Common biometric template Torn: a template Created In Vendor ps system ‘com be used through verdor 87s systern% Biornetric data suchas fingey prints and facial images cannot be eveconstructed From bromettic Leroplates + Biometrre templates ore generated every time osey presents biometric data To Successive Placement of a finger on ‘a biometric device Jenerate entirely diffevent templates “This is because of pesitionin distances ,pressuves ete Biomebic Matching: The comparision a biometric templates, to detew they Bogie of ssimilavity oy Correlation Ik called mamechng Scorin i A. number indicating the degree of Similarity OY Co- Welation Vesulting from the comparison of” enrolment and voracedlis ope Threshold - A threshold is a predefined number, generally choosen by a system “administrator, which establishes the dequee Of Oo-velation necessary for a core paaae ny to be beemed a match Decision : “the result Of the comparison between the Score and the threshold t¢ a decision Accuracy Of Biometric: Systems: The key performance matics in Biometrics Me false match vate Cee) jfalse Nen-Mmatch Rate Crwme) fouls to enroll vate (FTE)eee ’) Fatse’ Match Rate (fe) - fo ee ‘alse match vate is will be (6y & aiff A biometric solutions { obabitity that o Lsev template incovvec tly jedged to be a match users template False match may occur because (we people bave Cirmilay encugh. biometric. Chonacters tic . Tt is @ gatted as fatse acceptance vate £MR can be Yeduced by adjustin thrrechold thi Adjust thresholds that adjust Che level of corte CEion ae for ‘two templates & Jeelges a match 2) False Non-Malch Rate CFNMR) ‘alse Non J) x A. biometric sotvtion Eye ic the pyobabili that user template will Re incorrect joaged to Mot match vsers enyollment template a Authorized User IS locked out of a system, ine denied access t0 a facility ov ‘resource: Th ! alco calted false “rejection vate The factors affecting" FNMR are ) Changes in users bidbme tric dotte. ac cin how a user” presents biometric da 3) changes in the eo ORITERE in vohich date ot vvesented Failure to Enrollment: CETE)* Se D system failure 0 enroll vate sreprresents c Unable probability that a ae use y) will b enroll in a biometvic BeeFIE ts occuved tohen vsev’s have. insofficien td distinctive Oy ‘veplicoble biometyic data or when, the design of the bidmetric solution 1¢ such that providing” consistant data is difficult Derived Metrics “Theve ane ‘another two metrics Used to ve flect. the Overall occurancy ilities Of a) biometric technology These metrics-ane generated from Analysis and congpeediaan of FNMR, ETE DEqual Error Rate (EER): It is also catled as cyoss Ger vate -Tt fs the sate at which the FNMR is egUel to the FME ie, tt sepresents the aceirrasy level at which | anit ies cf a False match is the Same as the Ukelyhood of a false non-nmatch 9) Ability to verity Rate CATY) A move valued to devived metrics Is the ability to verify vate . ATV IS a combihation Of FTE and fNMe and Indicates the ovevall percento. e Of Dsers who will be capable of. authenticating On a daily basis aty = C- FTE) (1- FNM) This metric can be thought of as ead) O axcup of users eho CAPNEL eNvWll alon with users falsely ejected, by the Systeny tar pry vote will Makes for A move effective fnerect Biometvies toyered Biometric solutions ave these that ve the cubmission OF move than one biometric Chava for verification, such as finger Scan and voice ¢ or finger Scan and facial Scan. These solotions: a Seen as a method of weducing FMR [False Ma Ratel,eNMR (Fate Non Matching ate] (Envoy Rates “There ave two types oe sayered Biometrics 4) Parattel tayered Biometrics a) Sevial tayered Biometrics ¥ Pocrattel Layered Biometrics » Pocvallel Layered solutions wrequited that the submit mackple biometrics Ino. single Quthenti fal a Poss Match Finger scan oS 2. Decision Resubmit the - combined Facial scan Scove compared K with the ‘ Ue es threshold value two biornetric Choracterstice Simultanecust a: Serial Layered Biometrics : Gevial layered Solutions Combine the yesole two ov thee Sepevate authentication process authentication Users: Serial La med Use¥ intevackion might Consist OF verification on Fingerscan, thely facial scan then voice scangubrnit finger print 2. eons if fait L Fact image Paar Match vere dal foil No Matth ——» Resubmit ‘ the data Fingey Scan Technology Fingey scan Technology vtilises distinctive features oF the Finger Print to identify ov verify the individvale identi Fingey Scan systems Compyice Qequisi tion havd ware, image processing components, template "YAEION ~ and ‘matching components and storage components The surface on which finger 16 placed is called a platen also called as a ¢Scanney ae Depending on the type of Fingey scan technolog Osecd ecreas of contact between the finger print 09 the - platen by aes Unovgh Chip based cameras throu Ultrasonic i yt J cope fields Sere bet hey ot ae These Measurements are Converted to” ital Code x A platen attached to a small Printed circuit. boare ~ with a standard connector. that allows " digit ized wnformation to be trancmitted to the Pevipheval oy standatone deviceao oe Platen How finger scan seueaogy works : I Trage acquisition 2- Image Processin 2. Identification’ of distinctive chavacterstics q er Generation Ss Motehing tor. ; Scheipetic cola meemajeaape processing in final Sustems: Se f 1 evecetves the * Data si Device level a decision Fok Proces 23 “Transmitted Lectitas cessin anaes a Re ne a Server ae Scenanioa: Device level \ 2 Storage of + Data ¢ - template fox precesing and y Seah omy a : : Cocal PC 9 for ful storage a + Decision pe 9 trans mitives Scenarios docal Pe rnage 4 Pat prons and wdcgui- 2 “sition J storageScenaviog eimage central server IndiGitalized Datta. Stovedt 2 processing and ee ars 3 for vecord ing storage transmitted 4 forefigtone Ss * Im eI Pas ee “there are five stages involved tn Finger scan vevification and identification t-Fingey print imnage acquisition eemene processin 3. Location OF Distinctive chayactevsticc Y. Template ‘Creation ver ermplate Matchi * Each vendor's process differ slightly” Fron the sequence described TR sia 4-Emage Acq sision: Tmage quality 15 meacured in Dots Pex Inch (DPI) . Move dots for inch means a high Yecolution image - Today's Finger scan pert hevale Can acquire Images of S00 DPE the standaad Foventic- quality Fingey Printi Finge Print quality can vary * Strong Fingey Scan solutions ave capable of Processing thousand: of usexs withovt allow! Q false match and can .verie, nearly one hondred Percent OF users with one or Lwo Placements of a Finger * Many Finger scan.solotions Can be de Played ins Applications where both Seecnity and convenience are ag drives. C 2 Range 2F Deployment Environments : Reduced size and> power equivements along with Linger scan Yesistance to environmental chan ec Soch as background, Uk nang and temperato%e, Allow the techno: to be deployed ina eon of logicat and ysical juccess: envio There are more finger scar solutions in the. biometric market ‘plac than all othey stechnologies Combined 3 Easy _ to-ese devices: “The act of Placing a finer on a device i¢ largely inturkive and car be grasped tor &hLittle training. Many other biometric tec yequire comple x vser- System mtevac trons Scan devices ave Jeneralt designed such thi placement ts an easily repeatable Process y Ability to enroll pooltiple Fingers: ‘Most. people can envoll upto ten fingeys in@ biometric System gives finger scan advantage Security and Flexibility - Allowing a vser €0 verify with one oy several enrolled Fingers reduces the likelihood of a user being al rejected atti inget$ehb Se 1. Tnability +o |enyoll some osers : A small percent of users are unable to & in man Finger scan Systems Certain demo Groups have lower quatity Fingerprints and mi difficult to enroll them- Elderly opulation manv™ labourers and some asian lation ave MOVE likely. to be vnable to enroll in Some fae systems a-Pexformance Deterioration Over time ee =e Although the Finger print ts a Stable physiclo charvacterstic, dacty Wweay can Cause the perv of some fingex scan technologies to drop dias USexs Who Work With hands ave ese b se incveaced evyor yates Over time 3 . Assosiation twith forensic applications » finger Scan technotogies sini laity to forersie fingey printing causes Some excentage Bi ses discom forty Privacy advocates feay thatfingerscan data collected for a specific porpas may be used for forensic applications 4. Need to deploy specialized devices Tnorder for fingexscan sto become a. pervasive biometric olution) devices most be present on desktops, at points of sale, ab protected door way and at any Location where ounthentication ic vequited “Types oF algorithms used for interpretation: le need to Know which types of algorithms Ayre used to. know what constitutes a finger print ahd how rt cam be Imaged etc» the Following categories of algorithms are used, ANMinuti a- based algorithm aPattern- based algorithm 3) Hybrid - based algorithm DMinutia -based algorithnas Vendors who choose Eo use minutia based algorithms will need to provide the highest quatity ire possible Minutia. based templates ave relatively ematley than pattern base, templates -For minutiae baced finger print algorithm only a small part of the finger tmage is vequired for verification As such it 1s ideal to have ac much minutio ac Possible In the finger tmage avea- Since jost &@ ortion of the minutia. if Yequired’ foy verificate it would be ideal to use Enis algorithms (DBbye | space restriction impact the Use and: deployment of iniegnatsric-s A goed tmragey for @ minvtia -basced algoritrm would be one thai takes a high ualits image has a large enough capture window eo the velative cove o fing ev imdge to be imagedere a. Pattern - based algorithms : Pattern. based algovithms use both the mi macro features of a fingey print: When th macro features are utilised, the Size Of th Veguived fey authentication, when compayel the size of the Image needed for minutia, Tea uinements these pes OF algorithms ten, fast and have a larger template size - “They also mequired move of the image a ‘oe. present doving Verification +A good } for pattern - based matching algovithme tk that has a high quae Cameya anda enough me surfacé to capture the i macro ditailh. 4 2 +Hybrid- based Algonithm A bybrid algorithms uses the best features minutia based algorithms and those from pert based matchin. IS Algorithm is a good all agorithm giving QA Q00d bade off between acev1acy of the minvtia algorithm and Spe the pattern based ‘erecognition, A high Quote optical sensor fs best for these types of It woold offer a large PBA9e Image Area, very geek quality or the images e hybri algorithen takes longer to envotl becavse of vse cf both minvotia, and character based Yecognization» Once these has occured, t nesidled is) aetuatly factey than the mi based adiger/ittiisTemplate size is Wnportont +The yelative speed difference between Minvtia. based and pattern based Algorithme are. ie “Te Application doesnot requtve high through then minutia’ boue algorithm wovld work best : template Size is ‘Not impartant + ve lative speed difference is sufficiently meaning full «The op plication has high erro ae then = pattern based ao WOutA Work besy rE: u template size is not Im portant + faster matchi ts Fequiveel « the* Opptication has high eoToeeryaee thern > Hybrid based alaeri thm will work best 3