Secure Authentication Approach for Human using Finger Images

Secure Authentication Approach for Human using Finger Images

Ms. Monali S. Khonde

ME, Student, Dept. of Electronics & Telecommunication PVPIT, University of Pune,

Pune, India

Prof. S. M. Kulkarni

Dept. of Electronics & Telecommunication PVPIT, University of Pune,

Pune, India

Abstract Nowadays, authentication is a major concern in security. This paper presents a new approach to improve the performance of finger- vein identification systems. The sharing or theft of user passwords still remains the most frequent way that corporate data is compromised. This proposed system obtains the finger vein and finger texture and merge these two evidences using a novel score level combination. This paper examines previously proposed finger- vein identification approaches and develops a new approach to illustrate its superiority over prior published efforts. This new system also supports the utility of low resolution finger print images acquired from a camera is examined to determine the matching performance from such images. This paper investigates and develops two new score-level combination i.e. holistic and nonlinear fusions, and also comparatively evaluating them with other popular score-level approaches to establish their effectiveness in the proposed system.

Index TermsFinger-vein recognition, feature extraction, Personal Identification, Fusion (Holistic and Nonlinear).

I.INTRODUCTION

Automated human identification such as biometrics that uses human physiological or behavioral features for verification, has increasingly mapped for commercial use and also tremendously demands for more user-friendly and secured biometrics [1] has motivated researchers to explore new biometrics features and traits. For decades, the password has been widely considered as standard means for user verification on computers. However, as users are required to memorize and changing the passwords, it is essential that a more suitable and safe solution to user authentication is needed. For many researchers it is a challenge to find out a highly secured authentication technique using biometrics. So they have sought to improve the consistency and annoy spoolers by developing biometrics that are highly individuating [2]. The individuality of fingerprints has been recognized to the casual imperfections in the friction ridges and valleys. The acquisition of such minutiae features typically requires imaging resolution higher than 400 dpi. The convenience of such features acquired from the low-resolution (webcam) images or at a distance [4], deserves attention for its feasible use in personal identification for particular applications.

Some biometrics technologies are liable to satire attacks in which fake fingerprints and static face images can be effectively employed as biometric samples to imitate the identification. Hence, several liveness counter measures to detect such sensor-level spoof attacks have been proposed [1], [10], e.g., finger response to electrical impulse [5], finger temperature and electrocardiographic signals [6], time-varying perspiration patterns from fingertips [7], and a percentage of oxygen-saturated hemoglobin in the blood [8].

1. MOTIVATION AND RELATED WORK

   The principle behind the finger vein authentication technique is as follows: The pattern of blood vessels is caught by transmitting near infrared light and light is partially absorbed by hemoglobin in the veins and the pattern is caught by the webcam. The finger-vein patterns are believed to be quite unique, even in the case of identical twins. There are two factors that are cited for the preference of finger-vein biometrics. 1) The finger veins are concealed structures hence complicated to steal the finger-vein patterns of an any person without their awareness.2) It offers robust antispoofing capabilities as it can also guarantee liveness in the presented fingers during the imaging.

   Finger vein detection is a method that utilizes pattern- recognition techniques based on finger images of human finger vein patterns beneath the skin's surface of the finger to identify individuals identity. To obtain the pattern for the database record, the person has to insert a finger into an attester terminal including a near-infrared LED and a CCD camera. The CCD camera records the finger image and the raw data is digitized and sent to a database of stored registered images.

   1. PROPOSED SYSTEM

      The fingers presented for the identification of subjects are simultaneously exposed to a webcam and an infrared camera, as illustrated from the device of our imaging device in Fig. 2(a). The dorsal side of a finger is exposed to the near-infrared frontal

      surface illuminators, using light-emitting diodes whose illumination peaks are at a wavelength of 850 nm, whereas the frontal surface entirely remains in the contactless position with both of the imaging cameras. The finger-vein and finger texture images are simultaneously acquired using the switching device/ hardware that can switch the infrared illumination at a fast pace. The obtained finger-vein and finger texture images are first given to preprocessing steps, which automatically extract the region-of-interest (ROI) images while reducing the translational and rotational variations. The enhanced and normalized ROI images are employed to extract features and then generate matching scores similar with a conventional biometrics system.

      The combined matching scores are employed to authenticate the user.

      Image Enhancement

      Fig 1. Represents the block diagram of the system and its blockwise explanation is as given below:

      Image Normalization

      ROI

      Extractor

      Image Enhancement

      Finger image from database

      Finger Vein Image Preprocessing

      Image Normalization

      ROI

      Extractor

      Finger Texture Image Preprocessing

      1. FINGER-VEIN FEATURE EXTRACTION

         The normalized and enhanced finger-vein image from the imaging setup represents a vascular network with varying thickness, clarity, and uncertainty on the topological connections.

         Feature extraction

         Finger vein matching score

         Score Combination

         Enrollment database

         Imposter/ Genuine

         Feature extraction

         Finger Texture Matching score

         Fig 1. Block diagram of the proposed system

         1. FINGER-VEIN IMAGE PREPROCESSING

            The acquired finger images are noisy having rotational and translational variations resulting from unconstrained imaging. Therefore, the acquired images are first subjected to preprocessing steps where it includes 1) Image Normalization 2) ROI extraction 3) Image Enhancement

      2. FINGER TEXTURE IMAGE PREPROCESSING

         The obtained finger texture images from the webcam (640×480 pixels) are compact to 580 x380 pixels gray-level images because the cropped part does not supply any useful finger details. Then this image is employed for the preprocessing and get results.

      3. FINGER TEXTURE IMAGE FEATURE EXTRACTION

         The localized Radon transform (LRT) and the Gabor-filter- based extraction of such localized texture orientation details have shown to propose promising outcome for the identification of finger knuckles and palmprints.

         S {(Sv *) (St *(1))} {(Sv *) (St *(1))}

         (2 Sv)

         {(S *) (St *(1))}{(1

         1

         2 S )}

      4. GENERATING COMBINED MATCHING SCORES

      The combination strategy selected or developed should be robust and achieve significant performance improvement. These all output results can be combined at th feature, score, and decision levels.

      This approach searches for the linear combination that can provide the best separation of the genuine and imposter scores in the decision space and can be represented as follows:

      The final combined scores have a similar trend as the score from vein matching, when the score from finger-vein matching is high, the fused score will also become high and vice versa.

      • Nonlinear Fusion

        The nonlinear fusion score combination tries to dynamically adjust the combined score according to the degrees of uniformity between the two matching scores and is illustrated in the following:

        S ( c St ) * (c Sv)2

        c Sv

        S

        L

        j 1

        L

        (w j s j ), where w j 1

        j 1

        Where c is a positive constant and is fixed to 1 in this experiments and is selected in the range of [1, 2]. As the

        equation suggests, when the two matching scores are consistent,

        Where s j represents the scores from the jth matcher, wj denotes the weight that is assigned to the jth matcher (computed from the training data), and is the combined matching score.

   2. EXPERIMENTAL WORK AND RESULTS The real glory of finger vein authentication is its

      versatility. Because of its small and highly manageable size, it can be used for everything from gate-access control units to currency machines; car doors to PC log ins. In the future, it could even be used on portable electronic devices like cell phones and MP3 and also be adapted for identity documents like driving licenses, ID cards or passports.

      Here the two main approaches are used such as: First, a new approach for personal identification that utilizes simultaneously acquired finger-vein and finger surface (texture) images is presented. Second, this paper investigates two new score-level combination approaches, i.e., holistic and nonlinear fusion, for combining finger-vein and finger texture matching scores. We comparatively evaluate the proposed fusion approaches with the sum, average, product, weighted sum, likelihood ratio, and DempsterShafer fusion.

      .

      • . Holistic Fusion

      This approach is developed to utilize the earlier knowledge in the dynamic combination of matching scores. Let Sv , S t , and

      represent the matching score from finger vein, finger texture,

      and combined score, respectively, and this holistic rule of score combination is illustrated as follows:

      the final combined score is primarily contributed by the vein matching score.

      Finally after comparing both finger vein and finger texture it will generate the output as data stored in the database. It is shown by both software and hardware .The message will be displayed on the LCD display as match founds and simultaneously relay will be on the LED glows.

      Fig. 2 Circuit diagram of Hardware

      Fig. 2 shows the circuit diagram which is used as output. Where the output of software is given as input to the hardware relay circuit through RS232. If the vein and texture match founds then only the relay will be ON and message will be displayed on the LCD. If the match is not found then the relay will not get ON.

      The 10 database images are taken to generate the result on the basis of vein matching score, texture matching score, Equal error rate holistic, holistic fusion matching, Equal error rate non-linear, non linear fusion matching . All the statistics are shown in following table by precision.

      Sr	Vein	Texture	EEH	Holistic	EER	Non-
      .	Matchi	Matchi	R	Fusion	N	Linear
      no	ng	ng		Matchin		Fusion
      .	Score	Score		g		Matchin
      						g

      Fig.3 Table shows statistics of matching

   3. CONCLUSION

      This finger-vein matching technique works more efficiently in realistic scenarios and leads to a more precise performance, as demonstrated from the experimental results. Beyond embedded applications for portable IT devices such as cellular phones, finger vein authentication will in the future likely expand into applications such as opening automobile doors with a simple grip of the handle, for which the necessary grip-type authentication technology is already in development. Grip-type technology embeds personal authentication in the natural motion of opening a door, ensuring the highest security without forcing the user to learn complicated new procedures. This technology will be applicable to home, office or car doors and will usher in a secure future without keys.

   4. REFERENCES

1. Encyclopedia of Biometrics, S. Z. Li, Ed.. New York: Springer-Verlag, 2009.

2. D. Wang , J. Li, and G. Memik, User identification based on finger vein patterns for consumer electronics devices, IEEE Transactions on Consumer Electronics, vol. 56, no. 2, pp. 799-804, 2010.

3. A.K. Jain, Y. Chen, and M. Demirkus, Pores and ridges: High resolution fingerprint matching using level 3 features, IEEE Trans. Pattern Anal. Mach. Intell., vol. 29, no. 1, pp. 1527, Jan. 2007.

4. Handbook of Remote Biometrics, R. Chellappa,M. Tistarelli, and S. Z. Li, Eds. New York: Springer-Verlag, Jul. 2009.

5. P. Kallo, I. Kiss, A. Podmaniczky, and J. Talosi, Detector for recognizing the liveness character of a finger in fingerprint recognizing apparatus, U.S. Patent 6 175 641, Jan. 16, 2001.

6. D. Osten, H. M. Carin, M. R. Arneson, and B. L. Blan, Biometric personal authentication system, U.S. Patent 5 719 950, Feb. 17, 1998.

7. S. T. V. Parthasaradhi, R. Derakhshani, L. A. Hornak, and S. A. C. Schuckers, Time-series detection of perspiration as a liveness test in fingerprint devices, IEEE Trans. Syst., Man, Cybern. C, Appl. Rev., vol. 35, no. 3, pp. 335343, Aug. 2005.

8. P. V. Reddy, A. Kumar, S. M. K. Rahman, and T. S. Mundra, A new antispoofing approach for biometric devices, IEEE Trans. Biomed. Circuits Syst., vol. 2, no. 4, pp. 328337, Dec. 2008.

9. Finger Vein Products Aug. 2010 [Online]. Available: http://www.hitachi.co.jp/products/it/veinid/global/products/index.html

10. A. Antonelli, R. Capelli, D. Maio, and D. Maltoni, Fake finger detection by skin distortion analysis, IEEE Trans. Inf. Forensics Security, vol. 1, no. 3, pp. 360373, Sep. 2006.

    1	0.9299 62	0.17	1.211	98.7890 17	1.590 7	98.4093 28
    2	0.9286 09	0.21	1.239 6	98.7603 93	1.683 9	98.3161 31
    3	0.9227 82	0.29	1.291 3	98.7086 66	1.875 8	98.1242 22
    4	0.8696 24	0.23	1.156 8	98.8432 43	1.715 4	98.2846 02
    5	0.9194 0	0.23	1.224 9	98.7750 92	1.726 7	98.2733 16
    6	0.9249 62	0.22	1.241 1	98.7589 25	1.706 6	98.2933 74
    7	0.9069 62	0.21	1.202 8	98.7971 9	1.678 2	98.3218 35
    8	0.9146 99	0.31	1.292 8	98.7072 42	1.923 1	98.0769 46
    9	0.9257 89	0.17	1.203 9	98.7961 25	1.589 6	98.4103 61
    10	0.9186 84	0.25	1.253 4	98.7465 5	1.776 8	98.2231 61

11. Z. Liu, Y. Yin, H. Wang, S. Song, and Q. Li ,Finger vein recognition with manifold learning, Journal of Network and Computer Applications, vol.33, no.3, pp. 275-282, 2010.

12. E. C. Lee and K. R. Park, Restoration method of skin scattering blurred vein image for finger vein recognition, Electron. Lett., vol. 45, no. 21, pp. 10741076, Oct. 2009.

13. D. Mulyono and H. S. Jinn, A study of finger vein biometric for personal identification, in Proc. ISBAST, Islamabad, Pakistan, 2008, pp. 18.

14. L. Wang, G. Leedham, and S.-Y. Cho, Minutiae feature analysis for infrared hand vein pattern biometrics, Pattern Recognit., vol. 41, no. 3, pp. 920929, Mar. 2008.

15. A. Kumar and K. V. Prathyusha, Personal authentication using hand vein triangulation and knuckle shape, IEEE Trans. Image Process., vol. 18, no. 9, pp. 21272136, Sep. 2009.

16. J.-G.Wang,W.-Y. Yau, A. Suwandy, and E. Sung, Person recognition by palmprint and palm vein images based on Laplacianpalm representation, Pattern Recognit., vol. 41, no. , pp. 15311544, 2008.