Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/489504
Title: | Identification of spoofed fingerprint sample |
Researcher: | Akhilesh Verma |
Guide(s): | Savita Goel,Vijay Kumar Gupta |
Keywords: | Computer Science Computer Science and Engineering Computer Science Software Engineering Engineering and Technology |
University: | Dr. A.P.J. Abdul Kalam Technical University |
Completed Date: | 2023 |
Abstract: | newline Identification of individuals using biometric systems is now mature due to highlevels ofaccuracyinmatchingalgorithms,butitcomeswithasecurityriskofauthenticationon anindividual.Thisriskisknownasa PresentationAttack ,andthemethodsproposedby researchersforthedetectionofpresentationattacksareknownasFingerprintPresentation AttackDetection(FPAD).Thethesistitled IdentificationofSpoofedFingerprintSample attempts to highlight contributions made in the scholarly area of FPAD systems, propose a method for fingerprint spoof detection, and identify future researchdirections. newlineThe major limitations of the FPAD system are false-positive cases, where a spoof sampleisclassifiedasalivespecimen,whichisinherentinmachinelearningalgorithms. Inthepresenceoffalse-positivecases,atype-Ierroroccurs,affectingthefalseacceptance rate (FAR). The FAR is not desirable for operations in high-security risk environments, such as authenticated access to a banking system. Additionally, the increasing use of biometrics for general use by the common public is causing privacy risks. The end-user is well aware of the effects of spoofing attacks and how a copy of the biometric sample is made. Imposters are known to create fake biometric samples, and research labs have shown how to lift off or copy a biometric sample. Therefore, preserving the privacy of biometric data is unavoidable, and genuine authentication is uncertain by the FPAD system. Hence, technology needs to be competent enough to protect against presentation attacks, even if a biometric sample isstolen. newlineFPADsystemsarebroadlyclassifiedas Close-set and Open-set basedonlimitations and generalizationcapacity. newlineClose-set solutions are preferred in confined environments due to limited training cases. However, they cannot reproduce results in an open environment and come with issuesofaccountabilityandfairness.Adetailedreflectioniscarriedoutintheintroduction and literature section about all theseproblems. newlineInthebiometricindustry,serviceprovidersrelyonproprietaryclosed-setsolutionsfor |
Pagination: | |
URI: | http://hdl.handle.net/10603/489504 |
Appears in Departments: | Dean P.G.S.R |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
01_tittle.pdf | Attached File | 230.25 kB | Adobe PDF | View/Open |
02_prelim.pdf | 569.06 kB | Adobe PDF | View/Open | |
03_content.pdf | 490.38 kB | Adobe PDF | View/Open | |
04_abstract.pdf | 72.9 kB | Adobe PDF | View/Open | |
05_chapter_1_int.pdf | 1.95 MB | Adobe PDF | View/Open | |
06_chapter_2_lit.pdf | 4.43 MB | Adobe PDF | View/Open | |
07_chapter_3_met.pdf | 7.13 MB | Adobe PDF | View/Open | |
08_chapter_4_red.pdf | 5.52 MB | Adobe PDF | View/Open | |
09_chapter_5_fut.pdf | 1.35 MB | Adobe PDF | View/Open | |
10_annexures.pdf | 1.74 MB | Adobe PDF | View/Open | |
80_recommendation.pdf | 68.26 kB | Adobe PDF | View/Open |
Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).
Altmetric Badge: