a flc based fingerprint matching algorithm for images ... newly proposed fuzzy logic based...

Volume 1, No. 3, May 2012 ISSN – 2278-1080

The International Journal of Computer Science &

Applications (TIJCSA)

RESEARCH PAPER

Available Online at

http://www.journalofcomputerscience.com/

© 2012, http://www.journalofcomputerscience.com - TIJCSA All Rights Reserved

72

A FLC based Fingerprint Matching Algorithm for Images Captured with

Android Camera for enhanced Security of Online Transaction

Mangala R. Belkhede Computer Science & Engineering Department,

G.H.Raisoni College of Engineering, Nagpur, India.

[email protected]

Veena A. Gulhane Computer Science & Engineering Departmen,t

G.H.Raisoni College of Engineering, Nagpur, India.

[email protected]

Dr. P.R. Bajaj Electronics Engineering Department, G.H.Raisoni College of Engineering,

Nagpur, India. [email protected]

Abstract The next generation of banking applications won’t be on desktop or mainframes but on the small devices we carry every day. In this project authors have focused on, how biometric mechanism provides the highest security to the mobile payment. Secured e banking on the mobile is the latest issue for all mobile users. The present security issues surround the loss of personal information through the theft of the cell phone. The use of biometrics has been virtually eliminated the possibility of someone gaining access to a third party cell phone directly. It is therefore important that the biometric identification templates are not certainly stored on the phone but will gather at run time. The proposed biometrics mechanism secure the mobile payment also provides the security at the wireless transmission level. Mobile payment is used for banking and various M-commerce applications. Here authors are using the Android mobile for taking the real time fingerprint image for login the Mobile Banking Application. The main research focus on the online fingerprint matching algorithm for the run time finger images captured by 5 mega-pixel mobile camera for authentication. A newly proposed Fuzzy Logic Based fingerprint matching algorithms is implemented for matching the run time finger images.

Mangala R. Belkhede, Veena A. Gulhane, Dr. P.R. Bajaj, The International Journal of Computer Science & Applications (TIJCSA) ISSN – 2278-1080, Vol. 1 No.3 May 2012


73

Keywords- Biometric security, Mobile banking, Mobile Payment, Android, M-commerce

1. Introduction

The online banking transactions are part of daily routine for an individual. The existing online banking system has several drawbacks. Firstly hacking, from the internet anyone can hack the username and password and the result is third person gets access to owner account. As anyone is not with twenty four hours on the Internet, i.e. access bank website, it takes some time to know that your account get hacked and third one can get transfer the money to his own account. Secondly, every time one has to carry laptop or PC with you. So for this issue secured payment applications on mobile device i.e. M-commerce is proposed. Today is the era of mobile, everyone having the mobile in his hands, instead of using the laptop or PC, mobile is the best option to use for the banking purpose. The next generation of banking applications won’t be on desktops or mainframes but on the small mobile devices we carry every day. Secured e-banking on the mobile is the latest issue for all mobile users. M-commerce, in the context, provides a lot of services like Mobile ticketing, Mobile banking, Mobile location based services, Mobile auctions, Mobile purchasing and so on. This represents an incredible opportunity to enable mobile devices, as universal devices for mobile commerce applications. Existing smart phones in market an open, programmable software framework is vulnerable to typical smart phone attacks. Such attack can make the phone partially or fully unusable and cause unwanted SMS/MMS billing. The statistics shows [1,2] online transactions are hacked. Previously many online transactions are hacked. To avoid the general device attack, authors have used the Android mobile for the payment application. Android has software stack based on the Linux Kernel and it contain the Android Native libraries[3]. Android is very powerful device. As it having the in build libraries, and top level security mechanism to secure the rich application[4]. It also includes the Image processing library that can be used for the processing input images. PDAs and cell phones these days come with finger print scanners for authentication and transactions. There are various methods to take the runtime fingerprint. Android is having the inbuilt fingerprint scanner. It is also possible to install the fingerprint scanner software to the android device, and take the finger print at run time. Even if biometrics mobile is not available, the camera with high mega pixel can take the picture and can be processed further for the secured banking in android based mobile device. Here mobile digital camera is used to capture the fingerprint image. Finger print is a powerful mechanism in biometric authentication. So payment application is secured on all the ways, i.e. it uses the secure device, biometric security mechanism to open the payment application and wireless channel security.

2. Existing Work In a core banking system, there is a chance of encountering forged signature for



74

transaction. And in the net banking system, the password of customer may be hacked and misused. Thus security is still a challenge in these applications. There are many techniques to secure the customer information and to prevent the possible forgery of signatures and password hacking[2]. Still there are some problems. Today, single factor authentication, e.g. Passwords, is no longer considered secure in the internet and banking world. Easy-to-guess passwords, such as names and age, are easily discovered by automated password-collecting programs. Two factor authentications have recently been introduced to meet the demand of organizations for providing stronger authentication options to its users. A biometric technology makes sense for E-payment. A new approach to the fingerprint payment technology i. e. using biometric technology with E-payment is perfect because it won’t just identify, but it will authenticate as well. Dilip Kumar and Yeonseung Ryu have suggested to use fingerprint for operating ATMs, here in state of using the card, fingerprint get used for transaction[5]. The drawback/missing part in this paper that it was for ATM banking and not for handy operation with mobile or iPod dependent operation. Secondly Dr. Suresh Sankaranarayanan has worked on biometric mobile but these mobiles still are very expensive in the market hence technology cannot be available for common man, biometric scanner is used to take the fingerprint for the authentication on mobile device[6]. The authors have tried to simplify the process by using any mobile having 5 megapixel cameras. Biometric authentication technologies such as face, finger, hand, iris, and speaker recognition are commercially available today and are already in use [6],[7]. A biometric system is essentially a pattern recognition system that operates by acquiring biometric data from an individual, extracting a feature set from the acquired data, and comparing this feature set against the template set in the database. Depending on the context, a biometric system may operate either in verification mode or identification mode.

3. Methodology

3.1 Image Processing on Android

The mobile phone landscape change last year with the introduction of smart phone running Android, a platform marked by Google[8]. The Android operating system is preferable for benchmarking due to its recent growth in popularity with varying hardware manufactures e.g. Micro-max, HTC, Motorola, and Samsung. The Android operating system is supported and a part of the Open Handset Alliance. This alliance positions key manufacturers, cellular providers and the Android operating system in a collaborative environment which has caused large growth since October 2008 when the first Android mobile phone was released.

Fig 1: Image Processing Steps

Image processing on mobile phones is a new and exciting field with many challenges due to limited hardware and connectivity. Phones with cameras, powerful CPUs, and



75

memory storage devices are becoming increasingly common. The need for benchmarking basic image processing routines such as: addition, threshold and features extraction is important for comparison of systems. With this information developers and researchers can design complex computer vision and image processing applications while being aware of the current state of the art limitations and bottlenecks on mobile phones. Figure1 shows the basic image processing steps. Image Acquisition refers to the capturing of image data by a particular sensor or data repository. Once the image data is acquired, Pre-Processing often includes rendering the acquired data to a format that can be handled by a set of algorithms for Feature Extraction that transform sub image data to information which are often in turn maintained over time to provide temporal information. The main part is of focus on Image Acquisition and Pre-Processing through implementing image addition, threshold and features extraction on the mobile phone using the available Software Development Kit (SDK). The image processing library called JJIL (John’s Java Image Library) is used for its image decoding functions.[9] JJIL includes an image processing architecture and over 60 routines for various image processing tasks. It is particularly targeted towards mobile applications.

3.2. RGB Color Model

In color image, the color of a pixel is expressed by a RGB values in color space, there are different methods for color image processing and the gray image processing. RGB color space is used widely by image acquiring, RGB is one of basic color spaces. Traditionally, color images are represented in the RGB color space. RGB space, however, is not only a 3-dimensional space but also includes brightness or luminance which is not a reliable criterion for skin separation. It has been observed that finger colors of different people always be different, mostly due to intensity or luminance. However, luminance cannot be a reliable criterion for differentiating between regions of two images as it is affected by the varying ambient lighting conditions. Thus, we always prefer to deal with images in which luminance has been removed. In the RGB model , each color appears in its primary spectral components of red, green and blue. This model based on Cartesian coordinate system. Here all pixel values of R,G and B are assumed to be in the range[0,1]. The number of bits used to represent each pixel in RGB color space is called pixel depth. An RGB image in which each of the red, green and blue images is an 8 bit image. Under this condition each RGB color pixel is said to have a depth of 24 bits. An image is usually stored in RGB color space and the three components have correlation i.e. Red R, Green G and Blue B.[10] In Acquisition phase, image not having any background, all pixel are having the only finger image.

4. Research Design

Data flow diagram describes the flow of the project. The bank application is run on the linux system of Android. Run time fingerprint is captured by camera of Micromax A70 with Android operating system mobile for the login purpose of the bank application. Mobile will act as a client and the bank website will act as a server. Actually it is host server. Once fingerprint is taken as a login, it sent to the server for matching as request, and server send the reply message. If it is matching then only login will be successful and user can do the transaction. Fig 2 shows the data flow diagram of the project. Fig. 3 shows



76

the block diagram of the proposed biometric matching system. It is based on the RGB characteristics of the image and calculation of the RGB values of the image for matching.

Fig2 :Data Flow Diagram



77

Fig 3: Block Diagram of Proposed Fingerprint Matching System

5. Implementation

Fuzzy Logic Controller

Fuzzy Logic controller consists of a fuzzifier, rule based, inference engine and defuzzifier. In the fuzzy rule base, according to the problem’s requirements various rules are formed. Here numerical RGB input values to the fuzzier are converted into fuzzy values of register and testing images. These RGB fuzzy values along with the rule base are fed into the inference engine which produces RGB control values. At the defuzzifier the control values have to be converted to numerical output values. The block diagram of the fuzzy controller used in this paper is shown in Fig. 4.

Fig 4: Block Diagram of Fuzzy Controller for controlling RGB values for fingerprint

matching

The input parameter to this controller are the Red, Green and Blue pixel values of the register and test finger images. ( R1, G1, B1 are the values of register finger image and R2, G2, B2 are the values for the testing finger image.) Here authors consider the specific range of pixel values of thumb images. It will not consider the images above and below of specific ranges. Specific RGB values of human finger ranges are between 80 to 180 in RGB

Fuzzification The first input parameter considered in this controller is pixel value of red color. This measures with the 0 to 255. The pixel values are divided into the difference of 10. Red value of pixel between (0 to 10) -> 0, (11-20)-> 10, (21-30)-> 20, | |



78

(251-255)-> 250 According to this in the fuzzifier crisp values are stored. The second input parameter considered in this controller is pixel value of green color. This measures with the 0 to 255. The pixel values are divided into the difference of 10. Green value of pixel between (0 to 10) -> 0, (11-20)-> 10, (21-30)-> 20, | | (251-255)-> 250 According to this in the fuzzifier crisp values are stored. And the third input parameter considered in this controller is pixel value of blue color. This measures with the 0 to 255. The pixel values are divided into the difference of 10. Red value of pixel between (0 to 10) -> 0, (11-20)-> 10, (21-30)-> 20, | | (251-255)-> 250 According to this in the fuzzifier crisp values are stored. For considering the RGB values separately shifting function get used. a = (col & 0xff000000)>>24 r = (col & 0x00ff0000)>>16 g = (col & 0x0000ff00)>>8 b = (col & 0x000000ff) Inference Engine In the fuzzy Inference Engine, rule based are generated n n R1 =∑ (r) R2 =∑ (r) i=1 i=1 n n G1 =∑ (g) G2 =∑ (g) i=1 i=1 n n B1 =∑ (b) B2 =∑ (b) i=1 i=1 r, g, and b are the pixel values of red, green and blue respectively. n is the number of pixels in each row. R1 & R2 indicates red components for two different images after taking the average of each row. Similarly, G1 & G2 indicates green components for two different images, B1 & B2 indicates blue components for two different images. Now DR , DG and DB indicate the difference between red, green and blue components of two images respectively.



79

DR = R1 - R2, DG = G1 - G2 DB = B1 - B2,

If (DR >5) { Value=1} else { Value=0} Defuzzification Do for Green and blue factors % R= ∑ 0’s in DR /( ∑ 0’s in DR + ∑ 1’s in DR) % G= ∑ 0’s in DG /( ∑ 0’s in DG + ∑ 1’s in DG) % B= ∑ 0’s in DB /( ∑ 0’s in DB + ∑ 1’s in DB) %F = (R + G + B)/3 If %F is greator than 60% then Access Granted or Access Denied

6. Experimental Results

The Android mobile digital camera of 5 megapixel is used in the image acquisition. Images are captured by the 2x zoom function. Three samples of each user are stored at the database. Figure 5 shows the mysql database at the host server. Features vectors are stored according to the id of a respective client user. Average of each row of red, green and blue components are stored as a features vectors in the database. Proposed algorithm is implemented on the android SDK with the payment application. Figure 6 shows the running mode of finger print payment application on the android real device. Button “Click to Register” is used for registering new user and “Click to Check” is used for the login purpose. Figure 7 shows the Fingerprint Application with New User Registration. Here enter username and password for the registration, on clicking the “Register Button” camera activated and user get registered for the bank service. It also shows the uploading finger print image. Figure 8 shows the access denied i.e. testing fingerprint not matches with the database. Figure 9 shows the access granted so the transaction page has open and user can do the online transaction. Experiment result of proposed algorithm shows that it can achieve 90% accurate matching. Experiments are conducted by using total 30 user. Three samples of respective clients are stored at the database, and it gives the matching results. Statistically performance of the algorithm are shown in fig 10.

7. Conclusion

For the mobile payment we have being using till now information like credit card , signature and so on. These security mechanism are still not secure, so here we have introduce a biometric mechanism-fingerprint integrated with the camera for mobile payment system. Effective matching algorithm is implemented for the authentication with server database. Result will be shown through the screenshot. In future, multi-server authentication of the fingerprint can be implemented.

Possible Applications:

• Highly secured Mobile-banking



80

• Secured M-commerce

Future Implementation:

• Multi-server Authentication

Fig 5: Mysql Database Storing Features Vector Fig 6:Payment Application .

Fig 7:Fingerprint uploading (registration) Fig 8:Matching Results (Access Denied)



81

Fig 9:Matching Results (Access Granted) Fig 10:Performance of matching algorithm

References

[1] Han-Na You, Jae-Sik Lee, Jung-Jae Kim, Moon-Seog Jun, “A study on the two- channel authentication method which provides two-way uthentication in the Internet banking environment” [2] Chetana Hegde, Manu S, P Deepa Shenoy, Venugopal K R, L M Patnaik (2008) “Secure Authentication using Image Processing and Visual Cryptography for Banking Applications” [3] Asaf Shabtai,Yuval Fledel, Uri Kanonov, Yuval Elovici, Shlomi Dolev (2010), “Google Android: A Comprehensive Security Assessment.” IEEE security and Privacy. [4] Machigar Ongtang, Stephen McLaughlin, William Enck and Patrick McDaniel (2009) “Semantically Rich Application-Centric Security in Android” , Annual Computer Security Applications Conference. [5] Dilip Kumar, Yeonseung Ryu, “A Brief Introduction of Biometrics and fingerprint Payment Technology”, Published by the IEEE Computer Society,2008. [6] Dr Suresh Sankaranarayanan, “Biometric Security Mechanism in mobile Payment ”, Published by the IEEE Computer Society, 2010. [7] Fernando L. Podio, “Personal Authentication through Biometric Technology”, National Institude of Standards and Technology, Gaithersburg. [8] Margaret Butter (2011), Android: Changing the Mobile Landscape, Published by IEEE-ES 1536/1268/11

[9] Michael T. Wells ”Mobile Image Processing on the Google Phone with the Android Operating System”. [10] Advanced Digital Image Processing by Rafael C. Gonzalez (Author), Richard E. Woods (Author) [11] Nikunja K. Swain, “A Survey of Application of Fuzzy Logic in Intelligent Transportation Systems (ITS) and Rural ITS”- Southeast Con, Proceedings of IEEE, 2006, pp 85-89. [12] Dalal, S.; Satyanarayana, S. “Application of fuzzy logic to picture



82

quality improvement in televisions”, Digest of Technical Papers. ICCE, IEEE, 8-10 Jun 1993 Page(s):346 – 347. [13] S.J.Kher, P.R Bajaj, “Fuzzy control of head-light intensity of automobiles: design approach “proceedings of 37th SICE annual conference international session papers, July 1988, pp 1047-1050. [14] S.J. Kher, P.R.Bajaj, P.C.Sharma”A novel fuzzy control of headlight for night driving”, Intelligent vehicle symposium, proceedings of IEEE, 2000, pp-76-80. [15] A. V. Sai Balasubramanian, N. Ravi Shankar, S. Subbaraman, and R. Rengaraj ” A Novel Fuzzy Logic Based Controller to Adjust the Brightness of the Television Screen with Respect to Surrounding Light”, World Academy of Science, Engineering and Technology 39 2008,pg 188-191. [16] http://epsilon.nought.de/tutorials/fuzzy/fuzzy.pdf

a flc based fingerprint matching algorithm for images ... newly proposed fuzzy logic based...

Documents