ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: A Biometric Authentication Based Secured ATM Banking System Shouvik Biswas * Anamitra Bardhan Roy Kishore Ghosh Nilanjan Dey Department of CSE Department of CSE Department of CSE Department of IT JIS College of Engineering JIS College of Engineering JIS College of Engineering JIS College of Engineering Kalyani,WestBengal,India Kalyani,WestBengal,India Kalyani,WestBengal,India Kalyani,WestBengal,India shouvik089@gmail.com greatanamitra@yahoo.co.in com.kishore@gmail.com dey.nilanjan@ymail.com Abstract Authentication is a critical part of any trustworthy computing system which ensures that, only individuals with corroborated identities can log on to the system or access system resources. Biometric authentication with cryptography has ameliorated the security level for authentication. In this present work a Bio-metric system for encrypting the user password is proposed for secure remote ATM transaction. A retinal image is acquired at transaction terminal and Bio feature points are extracted from the blood vessel tree. User password is encrypted using some selective feature points. After transmission to the central server, encrypted password is decrypted using stored Bio-key to extract the password and check with the stored password. The success of the transaction is based on the stored database. The proposed scheme is self manipulative, simple, fast and yet much more secure. The efficacy of this computer simulation of bio-metric authentication system claims a secure online transaction. Keywords Color Planes, Genetic Operators, DWT, Harris Corner. I. INTRODUCTION There has been a significant surge in the use of Biometric based user authentication system in recent years because Biometric authentication offers several advantages over other authentication methods. Attacks when employed in securitycritical applications, especially in unattended remote applications such as e-commerce, it is very important to withstand. The present work aims at developing a novel crypto-bio authentication scheme in ATM banking systems [1]. At the time of transaction retinal image is acquired at the ATM terminal using high resolution Fundus camera. The blood vessel tree is extracted from the pre-processed retinal image. 256 bit Bio key is generated from the selective feature points extracted from the vessel tree using Harris Corner detection. User s password is encrypted by the proposed method using the bio-key. The encrypted password transmitted to the central server via secured channel. At the banking terminal the image is decrypted using the reversed steps of the same procedure with the help of stored Bio-key template. The success of authentication depends upon the matching result. Bio-metric based authentication can be a positive replacement of conventional password-based authentication. In this paper, Harris Corner [3] Detector is proposed to detect the corners present in retinal images as bio-metric feature points. The Harris corner detector is based on the local auto-correlation function of a signal; where the local autocorrelation function measures the local changes of the signal with patches shifted by a small amount in different directions. Based on the selective pixel s positional values of the corners a Bio key [4] is generated. This results into user specific Bio authentication key as biomedical features extracted from blood vessels tree, of every human eye is precisely unique. Hence these results into self authenticating characteristics of the image with high potential to recognize authorized relevant users. II. METHODOLOGY A. Discrete Wavelet Transformation The wavelet transform describes a multi-resolution decomposition process in terms of expansion of an image onto a set of wavelet basis functions. Discrete Wavelet Transformation has its own excellent space frequency localization property. Application of DWT[2] in 2D images corresponds to 2D filter image processing in each dimension. The input image is divided into 4 non-overlapping multiresolution sub-bands by the filters, namely LL1 (Approximation coefficients), LH1 (vertical details), HL1 (horizontal details) and HH1 (diagonal details). The sub-band (LL1) is processed further to obtain the next coarser scale of wavelet coefficients, until some final scale N is reached. When N is reached, 3N+1 sub-bands are obtained consisting of the multi-resolution sub-bands. Which are LLX and LHX, HLX and HHX where X ranges from 1 until N. Generally most of the Image energy is stored in the LLX sub-bands. 2012, IJARCSSE All Rights Reserved Page 178
III. PROPOSED METHOD Retinal Fundus image is collected from DRIVE database [10]. A. Key Generation from Retinal Fundus Image Figure 1. Three phase decomposition using DWT. The Haar wavelet is the simplest possible wavelet. Haar wavelet is not continuous, and therefore not differentiable. This property can, however, be an advantage for the analysis of signals with sudden transitions. B. Harris Corner Detection Harris corner detector [13,14] is based on the local autocorrelation function of a signal which measuring the local changes of the signal with patches shifted by a small amount in different directions. Given a shift ( x, y) to a point (x,y) the auto-correlation function is defined as : c(x,y) = w[i(xiyi)-i(xi+ x, yi+ y)]2 (1) Where I (xi, yi) represents the image function for (xi, yi) points in the window W cantered around (x, y). Here W is The Gaussian window is defined as where σ defines the width of the window. The shifted image is approximated by a Taylor expansion truncated to the first order terms: I(xi+ x, yi+ y) [I(xiyi)+[Ix(xiyi)Iy(xiyi)] (2) Where Ix (xi, yi) and Iy (xi, yi) indicate the partial derivatives with respect to xi and yi respectively. With a filter like [-1, 0, 1] and [-1, 0, 1] T, the partial derivates can be calculated from the image by substituting Eqn. (2) in Eqn. (1). C(x, y) the auto-correlation matrix captures the intensity structure of the local neighbourhood. For α1 and α2 be Eigen values of C(x, y), three cases may be considered as: 1. Both Eigen values are small signifying uniform region (constant intensity). 2. Both Eigen values are high signifying Interest point (corner) 3. One Eigen value is high signifying contour (edge) To find out the points of interest, Characterize corner response H(x, y) by Eigen values of C(x, y). C(x, y) is symmetric and positive definite that is α1 and α2 are >0 α1 α2 = det (C(x, y)) = AC B2 α1 + α2 = trace(c(x, y)) = A + C Harris suggested: the corner response H corner Response= α1 α2 0.04(α1 + α2)2 Finally, it is needed to find out corner points as local maxima of the corner response. Step1. Colored Retinal Fundus image is converted into gray image from the green channel. Step2. The gray image is pre-processed to extract Retinal blood vessel tree. Step3. The image is binarized and Binary area open is applied for removing the small objects. Step4. Sobel edge detection followed by image thinning is applied on binarized image. Step5. Harris Corner Detection is applied. Step6. All Harris corner detected pixel positions are computed. Step7. 8 x and 8 y co-ordinates are converted (decimal number) of first 8 Harris Corner detected position into 8 bit-binary number, which results a Bio-key of length 64. B. Password Encryption Using Bio-Key Step1: An 8x8 binary matrix is generated by converting the individual 8 bit Binarized ASCII values of each character of a 8 character password. Step2: The Bio-key matrix is decomposed into 16 groups each consisting of consecutive 4 bits row-wise and designated with numbers in order. Step3: Selected 4th, 5th, 6th, 7th bit of a row from the password matrix determines the corresponding decimal value to identify the corresponding group of Bio-key matrix. Step4: Hamming distance is calculated between the selected Bio key group and the 3rd, 4th, 5th, 6th bit of the selected password matrix row. Step5: 1st MSB is toggled; if the value of Hamming distance is even, else is left unaltered. Step6: Hamming distance is calculated between the selected Bio key group and the 5th, 6th, 7th, 8th bit of the selected password matrix row. Step7: 2nd MSB is toggled; if the value of Hamming distance is even, else is left unaltered. Step8: An encrypted matrix is generated by repeating Step 3 to 7 for each rows of the password matrix. Step9: Each row is converted into its corresponding decimal value ranging from 0 to 255. Step10: The encrypted key for all decimal values is generated, by considering the first 2 bits (ten & hundredth place) of the decimal value and extracting the English alphabet of that very position based on the decimal value leaving rest bits unchanged. Step11: The encrypted password is transmitted from the ATM terminal to the Central Server of the Bank over the channel. 2012, IJARCSSE All Rights Reserved Page 179
C. Password Encryption Using Bio-Key Step1: Based on the stored Bio key for the concerned user the Cipher text is decrypted applying the steps 2 to 10 in the reverse order to generate the original password. Step2: Original password is matched with respective password for the user in database of the bank and access is granted on successful matching, else declined. IV. RESULT AND DISCUSSION MATLAB 7.0.1 Software is extensively used for the study of the Biometric key generation process. Concerned images obtained in the result are shown in Fig. 2 through 7 and Bio operations are shown in Fig. 8 through 13. Figure 5. Extracted Retinal Blood Vessel Tree. Figure 6.Vessel Tree after Thinning. Figure 2. Color Retinal Fundus Image. Figure 7.Harris Corner Points. Total Count of Harris Corner is 791. The X co-ordinate of first 4 Harris Corner Detected points are 2,999,246,421 and Y coordinate are 2,2,426,434. A 8x8 Bio key Matrix is generated by converting these 4*2=8 decimal numbers into 8-bit binary numbers. Figure 3. Gray Image Extracted from Green Channel. 0 0 1 0 0 0 1 0 0 0 1 0 1 0 1 0 1 0 1 0 0 1 1 1 1 1 1 1 0 1 0 1 1 1 1 1 0 1 1 0 0 0 0 0 1 0 0 1 1 1 1 1 1 0 1 1 0 1 0 0 1 0 1 1 Figure 8. Generated Bio Key. Figure 4. Pre-processed Gray Image. 2012, IJARCSSE All Rights Reserved Page 180
Perform the same operation for all the rows of the password matrix. 0 0 1 0 1 1 1 1 1 0 1 0 1 0 0 1 1 0 1 0 0 0 1 1 1 0 1 1 0 0 0 1 1 1 1 1 0 0 0 0 Figure 9. Grouping of Bio key. User defined Password = Monti#10 (Size 8). Corresponding ASCII values are coveted into binary bits for individual characters of the given Password. M 77(ASCII value) 0 1 1 0 1 1 1 1 0 1 1 0 1 0 0 1 0 0 1 0 0 0 1 1 0 0 1 1 0 0 0 1 0 0 1 1 0 0 0 0 Figure 10. Generated Password Matrix. First row of the Password matrix is Selected. A B C Group the Password Matrix row based on the bit positions. 6th group( 0 0 0 1) of the Bio-key matrix is selected based upon the decimal value of selected Group A (0 1 1 0 6 ) of the selected 1st row of password matrix. Bitwise Hamming distance is calculated between Group 6 and Group B,Group C respectively. Group 6 Set B H.D of B Set C H.D. of C 0 0 0 1 1 1 0 1 1 0 1 1 0 0 1 1 1 0 1 0 Result of H.D. 2 5 The Hamming distance value of B is even, C is odd. So,only the,1st MSB value of the password matrix row is modified as: Figure 11. Modified Password Matrix. Consider only the first 2 bits of the decimal value. 255 as 25. 25 is Represented as Z based on the position value in the alphabet. Encrypted value of 255 is Z5. The generated string from this modified password matrix(encrypted Password) is U5E7L0L6Q9Q3R7Y0 Decryption Algorithm in applied on the following Matrix: 0 0 1 0 1 1 1 1 1 0 1 0 1 0 0 1 1 0 1 0 0 0 1 1 1 0 1 1 0 0 0 1 1 1 1 1 0 0 0 0 Figure 12. Decrypted Password Matrix. Applying matrix is: decryption algorithm the recovered password 0 1 1 0 1 1 1 1 0 1 1 0 1 0 0 1 0 0 1 0 0 0 1 1 0 0 1 1 0 0 0 1 0 0 1 1 0 0 0 0 Figure 13. Recovered Password Matrix. 2012, IJARCSSE All Rights Reserved Page 181
Recovered password is: Monti#10 At the banking terminal the password is decrypted using the same Stored Bio-key. The size of database reduces drastically as no image is stored. Also it reduces the complexity of transmission as only text files are transmitted. REFERENCE 1. A Method to Improve the Security Level of ATM Banking Systems Using AES Algorithm, N.Selvaraju, G.Sekar, International Journal of Computer Applications (0975 8887) Volume 3 No.6, June 2010 2. A Novel Approach of Image Encoding and Hiding using Spiral Scanning and Wavelet Based Alpha-Blending Technique, Nilanjan Dey, Sourav Samanta, Anamitra Bardhan Roy, IJCTA, Nilanjan Dey et al, Int. J. Comp. Tech. Appl., Vol 2 (6), 1970-1974 IJCTA NOV-DEC 2011, 1970, ISSN:2229-6093 3. A Comparative Study Between Moravec and Harris Corner Using Wavelet Based Image Fusion, Nilanjan Dey, Subhendu Das, International Journal of Advanced Research in Computer Science and Software Engineering, Volume 2, Issue 1, January 2012 ISSN: 2277 128X 4. David E. Goldberg. Genetic Algorithms in Search, Optimization and Machine Learning. Addison-Wesley, 1989 V. CONCLUSION In this work a Bio authentication scheme for ATM banking systems has been proposed. The retinal image of the user is required during a transaction. The password is encrypted via Bio key and then transmitted to the central server using Bio authentication. The encryption Bio keys are extracted from the Bio feature points of the retinal image, which absolutely user specific as per biological specificity. Any alphanumeric password with varying length can be encrypt by the method for transmission only by changing the the count of selected Harris points accordingly. 2012, IJARCSSE All Rights Reserved Page 182