A tm terminal design is based on fingerprint


Published on

for more projects visit @ www.nanocdac.com

1 Like
No Downloads
Total Views
On Slideshare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

A tm terminal design is based on fingerprint

  1. 1. ATM terminal design is based on fingerprint recognitionYun YangSchool of Electrical and Information EngineeringShaanxi University of Science & TechnologyXian, Chinayangyunll@163.comAbstract-For the traditional ATM terminal customerrecognition systems only rely on bank cards, passwords, andsuch identity verification methods which measures are notperfect and functions are too single. For solving the bugs oftraditional ones, the author designs a new ATM terminalcustomer recognition systems. The chip of S3C2440 is used forthe core of microprocessor in ARM9, furthermore, animproved enhancement algorithm of fingerprint imageincrease the security that customer use the ATM machine .Keywords-ATM terminal; ARM9; fingerprint recognition;image enhancement; GaborfilteringI. INTRODUCTIONWith the development of computer network technologyand e-commerce, the self-service banking system has gotextensive popularization with the characteristic offeringhigh-quality 24 hours service for customer. Nowadays, usingthe ATM (Automatic Teller Machine) which providecustomers with the convenient banknote trading is verycommon. However, the financial crime case rises repeatedlyin recent years, a lot of criminals tamper with the ATMterminal and steal users credit card and password by illegalmeans. Once users bank card is lost and the password isstolen, the criminal will draw all cash in the shortest time,which will bring enormous financial losses to customer. Howto carry on the valid identity to the customer becomes thefocus in current financial circle. Traditional ATM systemsauthenticate generally by using the credit card and thepassword, the method has some defects. Using credit cardand password can not verify the clients identity exactly. Inrecent years, the algorithm that the fingerprint recognitioncontinuously updated, which has offered new verificationmeans for us, the original password authentication methodcombined with the biometric identification technologyverify the clients identity better and achieve the purpose thatuse of ATM machines improve the safety effectively.II. THE CHARACTERISTICS OF THE SYSTEM DESIGNThe embedded ATM client authentication system isbased on fingerprint recognition which is designed afteranalyzed existed ATM system. The S3C2440 chip is used asthe core of these embedded system which is associated withthe technologies of fingerprint recognition and current high­speed network communication. The primary functions areshown as follows:• Fingerprint recognition: The masters fingerprintinformation was used as the standards of978-1-4244-6349-7/10/$26.00 ©2010 IEEE Vl-92JiaMiSchool of Electrical and Information EngineeringShaanxi University of Science and TechnologyXian, China10ckdotWia@yahoo.com.cnidentification. It must certify the feature of thehuman fingerprint before using ATM system.• Remote authentication: System can compare currentclients fingerprint information with remotefingerprint data server.• Telephone alarming: Once an exception happen,such as log in as the fake identity, the system willstart the phone alarm to inform client and bank staffas soon as possible.• Message alarming: the message can be send to therelevant staffs mobile phone without any noise, inorder to carry on emergency processing.• Police network connection: The system can call thepolice via the police network.• Two discriminant analysis methods: Besides thefingerprint recognition, the mode of passwordrecognition can be also used for the system.III. HARDWARE DESIGN AND SOFTWARE DESIGNThe design of entire system consisted of two part whichare hardware and software. The hardware are designed by therules of embedded system, and the steps of softwareconsisted of three parts. The more details are shown asfollows.A. Hardware DesignThe S3C2440 chip is used as the core of entire hardware.Furthermore, the modules of LCD, keyboard, alarm,fingerprint recognition are connected with the main chip(S3C2440).The SRAM and FLASH are also embodied in thesystem. There are some modules consisted of the system asfollows:• LCD module: The OMAP5910 is used in thismodule as a LCD controller, it supported 1024*1024image of 15 gray-scale or 3375 colors.• keyboard module: It can be used for inputtingpasswords.• Alarming module: TC35i alarming module is basedon GSM technology implement which can call thecredit card owner and send message to relevant staffswithout any sound.• SRAM and FLASH: The 16-bit 29LV160BB-70REC of FLASH chip and the 32-bitHY57V561620CT-6 of SRAM chip are connectedwith the main chip. Their functions are storing therunning code, the information of fingerprint and thealgorithm.
  2. 2. • Fingerprint recognition module: Atmel CompanysAT77CI04B be used as a fingerprint recognition. Ithas a 500dpi resolution, anti-press, anti-static, anti­corrosion.• Ethernet switch controller: RTL8308B can provideseight 10/100 Mbps RMII Ethernet ports, which canconnect police network and remote fingerprint dataserver.Before using the ATM terminal, the clients fingerprintfeature will be connected to the remote fingerprint dataserver to match fingerprint data with the masters, if theresult isnt correct, the system will call police automaticallyand send alarm to the credit card owner. The block diagramof hardware design is shown in figure 1.SRAMHY57v561620CT·6LCD moduleAT77CI04Bkeyboard module4*4S3C2440FLASH29LV160BB·70RFinger data serverFigure I. The block diagram of hardwareB. Software designThe design of software is very important for thisembedded system. The design was component of three partsincluded the design of main program flow chart, theinitializing ones, and the algorithm of fingerprint recognitionflow chart.This system of software is implemented by the steps asfollows: first of all, the Linux kernel and the File system areloaded into the main chip. The next, the system is initializedto implement specific task, such as checking ATM system,GSM communication and so on, and then each module resetfor ready to run commands.Before using ATM terminal, the password andfingerprint is required. First need to enter owners password,if password is successful then the system is required theowners fingerprint. If all the recognition is right, the systemwould enter into the waiting status. In addition, the numberof times that recognition of fingerprint and password arerestricted to 3.1f more than 3 times, the system will call thepolice through police network, telephone to the owner andsend message to relevant staff. Then locked the ownerscredit card. The overall flow chart of software is shown infigure 2.yyFigure 2. The overall flow chart of softwareIn the process of inputting fingerprint, the AT77CI04Bwhich is a linear sensor that captures fingerprint images bysweeping the finger over the sensing area, will used foracquiring the image of fingerprint. This product embed truehardware based 8-way navigation and click functions. Thefingerprint information will be temporarily stored in SRAMand upload to the remote fingerdata server to comparethrough bank network. The result of process will becontrolled by main chip(S3C2440).The initializing process means that set the hardware andsoftware and then start the multiple mission module, eachmodule will be started according to the priority processes. Atfirst, initialize the system clock, and execute the codes ofopen interrupt and the open interrupt task. Then, the systemwould judge and enter process of module. finally, the systemVl-93 2010 2nd International Conference on Computer Engineering and Technology [Volume 1]
  3. 3. would start to attempt multiple tasks. The initializing flowchart is shown in figure 3[ StartILIInitialize systemclockj;Open interrupttaskIiInitialize sy stemtaskj;]udge and enterprocess of modulej;Start to attemptmultiple tasksFigure 3. The flow chart of fingerprint recognitionC. The design offingerprint recognition algorithmThe design of algorithm based on fingerprint recognitionis so vital for the whole system. We would approach twosteps to process the images of fingerprint.1) The detail offingerprint recognition processThe first step was the acquisition of fingerprint image byabove device mentioned in the algorithm, and the resultscould be sent to the following process.Secondly, pre-processing the images acquired. Afterobtain the fingerprint image, it must be pre-processing.Generally, pre-processing of ones is filtering, histogramcomputing, image enhancement and image binarization.Lastly, the characteristic value was extracted, and theresults of the above measures would be compared with theinformation of owners fingerprint in the database so as toverify whether the character is matched, and then the systemreturned the results matched or not.2) The design offingerprint image enhancementFingerprint recognition module is an extremely importantpart of the system, the high-quality images was the majorfactors of influencing the performance in the system. Thereis a lot of noise in fingerprint image, the image enhancementwas the precondition for recognition of fingerprintcharacteristics. The algorithm of fingerprint recognitionbased on the algorithm of Gabor and direction filter wasused. fingerprint enhancement algorithm based on Gaborfilter could be better to remove noise, strengthen thedefinition between the ridge and valley, it couldsignificantly improve the image enhancement processingcapacity, but this algorithm was slow in dealing with thehigh capacity requirements. Fingerprint enhancementalgorithm based on direction filter has a faster processingcapabilities, but it was not good to handing the large noiseareas. so combination of these two algorithms could obtainbetter effects. The algorithm based on direction filter wasused in the clear area, and based on Gabor filter was used inthe recoverable region.a) The Gabor filter algorithmsFor each point, according to its frequency, direction,using formula(l) calculate the Gabor filter coefficients,calculate the filter values of each point and then ,move tothe next point, repeat the process above.hex,y,B,j)= exp -- X 2 +� cos(2Jifx) (1)II[ ,2 ,2J]2 8x 8yBesides,x=xsinB+ycosB (2)y=xcosB-ysinB (3)b) The directionfilter algorithmsFor each point, according to its frequency, direction,using formula (4) on the clear fingerprint image area.3 3h(x,y)= IIG(x+u,y+v)g(u,v) (4)u=-3v=-3G(u, v) is normalized after the fingerprint image, g (u, v)as a template filter coefficient.Proposed image block average M , the calculation usingimage block average gray level range, if the mean wassmall, the variance was small, that is not to restore area; ifthe mean moderate, have a great variance, the smaller theratio of mean and variance, it could be regard as the cleararea; if the mean moderate, variance and they were smallerthan the ratio of non-recovery area, it could be regard asconfusion region.Image block average calculated as follows:The fingerprint image was divided into w x w area,using equation (5), (6) in each of a gray-scale mean andvariance, And the statistics is greater than the number-- --of M ,marked as m and less than the number of M , markedas n.- {M-Bif 1 m-n I> w2, M= _M+Bm<n(5)m?nBis very small,1 W WM(n,m)=�IIh[H-(n-m)+u,H-(m-m)+v] (6)u=l v=1[Volume 1] 2010 2nd International Conference on Computer Engineering and Technology Vl-94
  4. 4. 1 w WN(n,m) =-2 LL{h[w(n-m)+u,w(m-m)+v]-M(n,m)}2(7)w u=1 v=1IV. CONCLUSIONSThe design of ATM tenninal system based on fingerprintrecognition took advantages of the stability and reliability offingerprint characteristics, a new biological technology basedon the image enhancement algorithm of Gabor and directionfilter. Additional, the system also contains the originalverifying methods which was inputting owners password.The security features were enhanced largely for the stabilityand reliability of owner recognition. The whole system wasbuild on the technology of embedded system which makesthe system more safe, reliable and easy to use.ACKNOWLEDGMENTThis project was supported by the Graduate InnovationFund of Shaanxi University of Science and Technology.REFERENCES[1] Lin Hong, Wan Yifei, Anil Jain. Fingerprint image enhancement:algorithm and performance evaluation[J]. IEEE Transactions onPattern Analysis and Machine intelligence. 1998,20(8): 777-789.[2] ESaatci, V Tavsanogh. Fingerprint image enhancement using CNNgabor-Cpe filter[C]. Proceedings of the 7th IEEE InternationalWorkshop on Cellular Neural Networks and their Applications 2002:377-382.[3] Gu J,Zhou J,Zhang D.A combination model for orientation field offingerprints. Pattern Recognition,2004,37: 543-553.[4] Cheng J, Tian J. Fingerprint enhancement with dyadic scale-space.Pattern Recognition Letters,2004,25(11): 1273-1284.[5] Chen H, Tian J. A fingerprint matching algorithm with registrationpattern inspection. Journal of Software,2005,16(6): 1046-105.[6] Smits G FJordaan E M.Improved SVMRegression using Mixtures ofKernels[A]. Proceedings of the 2002 International Joint ConferenceonNeural Networks[C]. Hawaii: IEEE. 2002. 2785-2.[7] FU Zhenghua, LI Yongjun, Tian Mi(2007). "The embeddedmonitoring system based ARM". JOURNAL OF 7INSTRUMENTTECHNOLOGY,Vol. 07,No. Ipp. 01-2.[8] Jun Zhou, Guangda Sua, Chun hongJiang. A face and fingerprintidentity authentication system based on multi-route detection.Neurocomputing 70 (2007)922-931.[9] Yuliang He, Jie Tian, Xiping Luo, Tanghui Zhang.Imageenhancement and minutiae matching in fingerprint verification.Pattern Recognition Letters 24 (2003)1349-1360.[10] Wei Wang, Jianwei Li, Feifei Huang, Hailiang Feng. Design andimplementation of Log-Gabor filter in fingerprint imageenhancement.Pattern Recognition Letters 29 (2008)301-308.Vl-95 2010 2nd International Conference on Computer Engineering and Technology [Volume 1]