Unconstrained fingerprint recognition fully touchless 3 d system ieee
Download as PPTX, PDF2 likes243 views
The document discusses a fully touchless 3D fingerprint recognition system designed to address the limitations of traditional contact-based methods. It outlines the proposed system's process, which includes contactless acquisition, 3D reconstruction, and matching techniques, showcasing experimental results that demonstrate improved accuracy and user acceptance compared to existing technologies. The conclusion emphasizes the benefits of touchless systems in terms of usability, robustness to environmental factors, and operational constraints.
![REFERENCES
[[1] D. Maltoni, D. Maio, A. K. Jain, and S. Prabhakar, Handbook of
Fingerprint Recognition, 2nd ed. Springer Publishing Company,
Incorporated, 2009.
[2] R. Donida Labati and F. Scotti, “Fingerprint,” in Encyclopedia of
Cryptography and Security (2nd ed.), H. van Tilborg and S. Jajodia,
Eds. Springer, 2011, pp. 460 – 465.
[3] R. Donida Labati, V. Piuri, and F. Scotti, Touchless Fingerprint
Biometrics, ser. Series in Security, Privacy and Trust. CRC Press,
2015.
[4] G. Parziale, “Touchless fingerprinting technology,” in Advances in
Biometrics, N. K. Ratha and V. Govindaraju, Eds. Springer London,
2008, pp. 25–48.
31](https://image.slidesharecdn.com/unconstrainedfingerprintrecognitionfullytouchless3dsystemieee-160928085140/85/Unconstrained-fingerprint-recognition-fully-touchless-3-d-system-ieee-31-320.jpg)
Unconstrained fingerprint recognition fully touchless 3 d system ieee
- 1. Toward Unconstrained Fingerprint Recognition: a Fully Touchless 3-D System Based on Two Views on the Move By, Sinisha George S2 MTech CS
- 2. OVERVIEW • Problem Statement • Introduction • Related works • Proposed method • Experimental Results • Conclusion • References 2
- 3. PROBLEM STATEMENT • Require contact of the finger with any acquisition surface • Require constrained and highly cooperative acquisition methods. • Low usability, user acceptance and presence of distortions, less robust to dust and dirt , hygiene issues. 3
- 4. INTRODUCTION • Fingerprints are one of many forms of biometrics used to identify individuals and verify their identity. • fully touchless fingerprint recognition system based on the computation of three-dimensional models. • Models from two view image captured during movement • less-constrained biometrics aim at using samples captured – Contactless, higher distance, Natural light conditions, On the move 4
- 5. RELATED WORKS A) Two dimensional system 5 Structured light
- 6. RELATED WORKS • Three dimensional system 6
- 7. PROPOSED SYSTEM BIOMETRIC RECOGNITION PROCESS 1) Acquisition 2) Three-dimensional fingerprint reconstruction 3) Computation of touch-compatible images 4) Template computation 5) Matching 7
- 10. CONTACTLESS ACQISITION(Cont.) Fig : Rotations of the finger with respect to the optical centre of Camera A. 10
- 11. CONTACTLESS ACQISITION(Cont.) Examples of images in Dataset B captured with exaggerate finger orientations: (a) leftward yaw rotation; (b) counter clockwise roll rotation; (c)downward pitch rotation; (d) rightward yaw rotation; (e) clockwise roll rotation; (f) upward pitch rotation. 11
- 12. 2. THREE-DIMENSIONAL FINGERPRINT RECONSTRUCTION 1) Image pre-processing; 2) Segmentation; 3) Extraction and matching of the reference points; 4) Refinement of the pairs of corresponding points; 5) Three-dimensional surface computation and image wrapping. 12
- 13. 2.1 IMAGE PREPROCESSING • Details of the ridge pattern are Particularly visible in the green channel of the RGB color Space of the captured images 13
- 15. 2.3 EXTRACTION AND MATCHING OF THE REFERENCE POINTS 15
- 16. 2.3 EXTRACTION AND MATCHING OF THE REFERENCE POINTS(CONT.) 16
- 17. 2.4 REFINEMENT OF THE PAIRS OF CORRESPONDING POINTS • To obtain smooth and accurate representation of the finger surface • Applying thin plate spline to the set of corresponding points. 17
- 18. 2.5 THREE-DIMENSIONAL SURFACE COMPUTATION AND IMAGE WRAPPING. • Creates a three-dimensional model MA as the depth map corresponding to the view of Camera A. • Triangulation • Linear interpolation – Mp, Mz Fig: dense three-dimensional model with a superimposed texture image obtained by the proposed method 18
- 19. 3. COMPUTATION OF TOUCH-COMPATIBLE IMAGES • Enhancement o background subtraction o Nonlinear equalization(logarithm) o Butterworth low-pass filter • Two dimensional mapping – Enrolment oCompensate for rotations oComputation of NR rotations 19
- 20. 3. COMPUTATION OF TOUCH-COMPATIBLE IMAGES(cont.) 20
- 21. 4.TEMPLATE COMPUTATION • Neurotechnology veriFinger – Commercial feature extractor – Compute template T – Properly identifies the coordinates – Designed for touch-based images. Fig: a binary image and minutia coordinates obtained using the proposed system and the software Verifinger 21
- 22. 5. MATCHING • In the verification phase computes a single minutiae template Tf . • Matching score represents the similarity of the fresh template with a multi-template. • Te score is computed as, • Match(.) – matching function ,performed using nuerotechnology verifinger software 22
- 23. EXPERIMENTAL RESULTS a) Data sets description b) Accuracy of 3D reconstruction c) Recognition performance d) User acceptability e) Interoperability 23
- 25. A) DATA SETS DESCRIPTION Touchless - one session. 2368 samples. 10 fingers, 30 volunteers, 8 acquisitions per finger. Touchless - two sessions 2368 samples 10 fingers, 15 volunteers, 16 acquisitions per finger. - 8acquisitions one year, 8 acquisition subsequent year . Touchless - misplaced fingers 1200 samples 2 fingers (index), 30 volunteers, 20 acquisitions per finger Touch-based One session Two sessions 25
- 26. B) ACCURACY OF 3D RECONSTRUCTION • point clouds describing the three-dimensional shape of the finger and the corresponding dense three-dimensional models with superimposed texture images • Average error 0.03mm 26
- 27. C) RECOGNITION PERFORMANCE • ROC curves representing the accuracy of the proposed touchless system under standard operating conditions (Dataset). •The results represent different numbers of three-dimensional rotations NR performed during the enrolment step. • Every test included 5, 605, 056 identity comparisons. •The configuration that yielded the best accuracy was NR = 9, with EER = 0.03%. Table: Accuracy of the proposed biometric system using samples acquired under standard operating conditions (dataset A).27
- 28. D) USER ACCEPTABILITY • Survey performed using questionnaires. • Results show preference towards contactless recognition. 28
- 29. G) OVERVIEW OF THE DIFFERENT TECHNOLOGIES 29 Aspect Touch -based Touchless accuracy EER=0.06% EER=0.03% Usability medium High User acceptance Medium High Privacy Data protection techniques Data protection techniques Speed Template extraction + matching 3D reconstruction + template extraction + matching Cost 10$ to 5000$ 10$ to 5000$
- 30. CONCLUSIONS – Systems based on two-dimensional samples can be used in low-cost applications, but the samples present distortions. – Systems based on three-dimensional samples can obtain comparable accuracy with respect to traditional systems – Touchless systems are characterized by higher usability, user acceptance, less constrained. – Robust to uncontrolled environmental illumination. – Tolerate wide range of finger orientations. 30
- 31. REFERENCES [[1] D. Maltoni, D. Maio, A. K. Jain, and S. Prabhakar, Handbook of Fingerprint Recognition, 2nd ed. Springer Publishing Company, Incorporated, 2009. [2] R. Donida Labati and F. Scotti, “Fingerprint,” in Encyclopedia of Cryptography and Security (2nd ed.), H. van Tilborg and S. Jajodia, Eds. Springer, 2011, pp. 460 – 465. [3] R. Donida Labati, V. Piuri, and F. Scotti, Touchless Fingerprint Biometrics, ser. Series in Security, Privacy and Trust. CRC Press, 2015. [4] G. Parziale, “Touchless fingerprinting technology,” in Advances in Biometrics, N. K. Ratha and V. Govindaraju, Eds. Springer London, 2008, pp. 25–48. 31
- 33. 33
- 35. Minutiae 35
- 36. 36
- 42. Morphological opening operator Morphological filling
- 44. Epipolar lines
Editor's Notes
- #4: 1. Two diamenional systems ::::: Doesn’t compnaste for rotations Presence of distortions Less robust to dust and dirt Less user acceptance hYgine issues Fingerprint recognition or fingerprintauthentication refers to the automated method of verifying a match between two humanfingerprints. Fingerprints are one of many forms of biometrics used to identify individuals and verify their identity. Biometrics refers to metrics related to human characteristics. Biometrics authentication (or realistic authentication) [note 1] is used in computer science as a form of identification and access control. It is also used to identify individuals in groups that are under surveillanc
- #5: Oon movement-so less user coorpration is needed System is less constrained tan 3d systems Fingerprint recognition or fingerprintauthentication refers to the automated method of verifying a match between two humanfingerprints. Fingerprints are one of many forms of biometrics used to identify individuals and verify their identity. Biometrics refers to metrics related to human characteristics. Biometrics authentication (or realistic authentication) [note 1] is used in computer science as a form of identification and access control. It is also used to identify individuals in groups that are under surveillanc
- #7: This also Touc base system but uses multiple cameras therefor 3d 5 cameras and set lED ligts
- #13: The proposed method computes a three-dimensional model of the finger with a superimposed texture representing the ridge pattern. The method uses a correlation-based technique that is more computationally efficient and accurate
- #14: Fig. 5. Example of a fingerprint image before and after the preprocessing step: (a) input image I; (b) ridge pattern image P, computed as the green channel of I. Because a green light is used to illuminate the finger. Consequently, we consider the channel G of the captured image I (Fig. 5 a) as the matrix representing the ridge pattern P
- #15: 2 task rough estimation of te finger sape Finger nail removal compensation of yaw angle search of finger rermination skin region segmentation
- #18: Fig. 9 presents examples of point clouds obtained with and without the refinement step.
- #19: f is the focal length of the two cameras, T is the baseline distance between the two cameras, xA and xB are the two matched points.
- #27: Fig. 16. Examples of point clouds describing the three-dimensional shape of the finger and the corresponding dense three-dimensional models with superimposed texture images: (a, c, e) point clouds; (b, d, f) dense three-dimensional models. The experiments demonstrated that the proposed method can obtain accurate three-dimensional reconstructions for all 10 fingers.
- #28: The EER is the best single description of the Error Rate of an algorithm and as lower be the EER the lower error rate of the algorithm- EQUAL ERROR RATE False match rate (FMR, also called FAR = False Accept Rate): the probability that the system incorrectly matches the input pattern to a non-matching template in the database It measures the percent of invalid inputs that are incorrectly accepted