Pith. sign in

REVIEW 4 major objections 6 minor 64 references

A Framework for the Security and Privacy of Biometric System Constructions under Defined Computational Assumptions

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper argues that biometric verification and identification can be formalized as computational problems and that universal composability lets their security proofs be modular.

desk verdict The abstract promises a UC-based biometric security framework, but the body never defines it: no ideal functionality, no simulator, no security notion, and the two lemmas are tautological, with Section 6 a literal copy-paste of Section 5. read the letter →

arxiv 2411.17321 v1 pith:GQU3PRZS submitted 2024-11-26 cs.CR

Sam Grierson , William J Buchanan , Craig Thomson , Baraq Galeb , Chris Eckl This is my paper · ORCID
classification cs.CR
keywords biometricsecurityuniversalcomposabilityidealfunctionalityverificationidentificationmetricspacesPAClearningprivacy
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that biometric system security and privacy can be proved rigorously rather than argued heuristically, by treating verification and identification as computational problems and analysing constructions through universal composability. It defines the biometric verification problem as a computational decision problem and the identification problem as a computational search problem, then presents protocol constructions with init, enroll, and verify subroutines that map biometric samples into a metric space via a learned embedding $g = h \circ f$. The central lemmas state that these constructions solve the corresponding problems for any metric space and mapping $h$. If the framework works, a system integrator could prove the security of the whole biometric system by proving each component's behaviour, with privacy and correctness grounded in explicit computational assumptions. The promised UC ideal functionality is the vehicle for that modularity.

What carries the argument

The machinery is the ideal functionality of a biometric authentication system under universal composability, a cryptographic paradigm in which a real protocol is compared with a trusted ideal process so that proving each component indistinguishable from its ideal specification yields security for the composition. The other load-bearing piece is the feature mapping $g = h \circ f$, where $f$ is a feature extractor from raw biometrics $X$ to a feature space $F$ and $h$ is a hypothesis, assumed to be the PAC hypothesis, embedding $F$ into a metric space $(M,\Delta)$; the matcher then accepts when $\Delta(g(x'), g(x)) \le t(\lambda)$. Constructions 5.1 and 6.1 are the protocols this machinery is supposed to analyse.

What would settle it

Take any concrete instantiation of Construction 5.1 with a specified metric space, mapping $h$, and threshold $t$, and measure $\Pr[\Delta(g(x'),g(x)) \le t(\lambda)]$ over the enrolment and challenge distributions; the claim requires this probability to be at least $1-\varepsilon(\lambda)$ for negligible $\varepsilon$, so a dataset where it is bounded away from 1 refutes the construction's correctness.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is a formal reduction: the biometric verification problem and the biometric identification problem can be posed as computational problems, and the standard four-module biometric architecture can be captured by a protocol Construction 5.1 (verification) and Construction 6.1 (identification) whose correctness follows from the properties of a metric space and a PAC-learned embedding into it. The paper states Lemma 5.1 and Lemma 6.1 asserting that the constructions are solutions, and it frames the contribution as the definition of an ideal functionality for biometric authentication under universal composability, from which security and privacy of the whole system are supposed to follow. In other words, the paper's claim is that 'secure biometric system' can be turned from a heuristic phrase into a proof obligation decomposed across components.

Load-bearing premise

The whole framework rests on the assumption that a good-enough learned embedding from biometric samples into a metric space exists, but the paper never specifies the data distribution, the allowed error, or how the embedding is trained.

Editorial extensions

If this is right

  • A security proof for a biometric system would no longer be monolithic; each module (sensor, extractor, matcher, database) could be certified separately and composed.
  • The same framework covers both verification and identification, so results proven for one mode transfer structurally to the other.
  • Instantiations in the Hamming metric (fingerprint-like binary codes) and the Euclidean metric (face-like embeddings) become candidate provable systems rather than purely heuristic ones.
  • Privacy properties, such as what an adversary can learn from stored templates, become expressible as ideal-functionality requirements instead of informal design goals.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial extension: the abstract promises the UC ideal functionality and its security proof, but the text available here contains the problem formulations and constructions without the UC definitions or the proofs of Lemmas 5.1 and 6.1, so a reader cannot yet verify the central security claim from this version.
  • Editorial extension: a concrete next step would be to instantiate the construction with an explicit feature extractor and PAC bound, such as a trained network with a measured generalisation error, to derive a concrete $\varepsilon(\lambda)$ for a real dataset.
  • Editorial extension: the framework's promise implies that template-protection techniques such as fuzzy extractors could be slotted in as components and composed, but the paper does not yet show such a composition.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper claims to introduce a formal framework for the security and privacy of biometric system constructions, with the abstract stating that universal composability (UC) is leveraged to derive strong security and privacy properties under well-defined computational assumptions. The body provides background on metric spaces, PAC learning, and neural networks; formulates biometric verification and identification as probabilistic decision/search problems; and presents two constructions with stated lemmas that the constructions "solve" these problems. The conclusions reiterate that an ideal functionality for a biometric authentication system has been defined and used for UC-based security proofs.

Significance. If the advertised framework existed, it could be a useful contribution: rigorous, composable security definitions are indeed missing from much applied biometrics literature, and a modular treatment of enrollment, verification, and identification would be valuable. The paper does contain a helpful survey of biometric system components and performance metrics (Section 3), and it correctly identifies the FMR/FNMR trade-off. However, the central deliverable is absent: no ideal functionality, simulator, environment, or security experiment is defined anywhere in the manuscript, and no computational assumption is ever stated. The central lemmas are asserted without proof, and one of them is essentially a restatement of the problem definition. The paper also contains a near-verbatim duplication of Section 5 in Section 6, with the copied text still referring to "biometric verification" inside the identification section. As a result, the claimed UC-based security and privacy framework cannot be evaluated, and the manuscript does not support its title or abstract.

major comments (4)
  1. [Abstract and §1 (Contribution); §7] The paper's central claim is that universal composability is used to derive security and privacy properties from defined computational assumptions. No ideal functionality, simulator, environment, or UC security notion is defined anywhere in the manuscript. There is no security experiment, no adversary model, and no computational assumption (e.g., hardness of a primitive) is stated. Section 7 merely asserts that the framework "uses the concept of an ideal functionality" without providing it. The load-bearing deliverable promised in the abstract is therefore missing, and the central claim is unformulable from the submitted text.
  2. [Definition 5.1, Construction 5.1, Lemma 5.1] Lemma 5.1 states that Construction 5.1 is a solution to the biometric verification problem, but no proof is given. More seriously, the lemma is circular: Definition 5.1 defines success as the probability that ∆(g(x'), x) ≤ t(λ), and the verify subroutine of Construction 5.1 returns exactly 1{∆(g(x),y)≤t(λ)}. The lemma therefore proves only that the construction's acceptance predicate matches the problem's acceptance predicate. This is a correctness condition, not a security or privacy statement, and it does not justify the paper's claims about UC security.
  3. [Definition 6.2, Construction 6.1, Lemma 6.1] Section 6 is a near-verbatim copy of Section 5. Definition 6.2 is identical to Definition 5.1 except that it still calls the problem "the biometric verification problem" in its opening sentence, even though it is presented as the identification problem. Construction 6.1 has the same init, enroll, and verify subroutines as Construction 5.1, and it does not implement one-to-many identification: there is no search over the database, no argmax, and no possibility of returning ⊥ when no template is close enough. Lemma 6.1 thus cannot be true for the stated identification problem, and the identification contribution is not a separate construction.
  4. [Construction 5.1, init subroutine; §2.2] The correctness of both constructions depends on the assumption, stated in init, that "h is the PAC hypothesis" for mapping the feature space to the metric space. No distribution over biometric data, no error bound, no risk function, and no training procedure are defined, and no PAC learnability result is proved or cited that would connect the existence of such h to the success probability in Definition 5.1. If such an h does not exist or has unacceptably high error, the verification probability is not guaranteed. This is an unstated, load-bearing assumption that the paper does not discharge.
minor comments (6)
  1. [Title and front matter] The title contains a typo: "PRIV ACY" should be "PRIVACY."
  2. [Definition 5.1] The probability event is written as ∆(g(x'), x) ≤ t(λ), but x is in X while g(x') is in the metric space M; the intended comparison is presumably ∆(g(x'), g(x)) ≤ t(λ). This type mismatch makes the definition formally incoherent.
  3. [§2.2 (PAC Learnability Framework)] The PAC subsection ends with unlabeled equations and no explanatory text; the definitions of ErrD, the ERM objective, and the expected loss are presented without connecting prose or a formal PAC learnability definition.
  4. [Construction 5.1, init subroutine] The init subroutine lists "EER" as a parameter to be defined by V, but no value or definition is supplied, and the equal error rate is not used in any later subroutine.
  5. [Example 6.1] Example 6.1 is titled "Face Verification in the Euclidean Metric" although it appears in the identification section; either the title or the placement is inconsistent with the intended content.
  6. [References] Many references are incomplete: some entries lack year, venue, or page numbers (e.g., [2], [6], [7], [29], [30], [51]), and several URLs are informal or outdated. The reference list requires a thorough cleanup.

Circularity Check

2 steps flagged · score 8.0 of 10

Lemma 5.1 and Lemma 6.1 are satisfied by definition: the verify subroutine returns exactly the threshold predicate used to define the problem, and Section 6 copies Section 5 verbatim under the label 'identification'; the promised UC/ideal-functionality derivation is absent from the text.

  1. self definitional [Section 5.1, Definition 5.1, Construction 5.1, Lemma 5.1]
    "Definition 5.1: "Pr[ ∆(g(x′), x) ≤ t(λ) | D := {(idi, g(xi))} ℓ(λ) i=1 ... ] ≥ 1 − ε(λ)" ... Construction 5.1: "verifypp(id, x) → b: V returns b := 1{∆(g(x),y)≤t(λ)}" ... Lemma 5.1: "For a metric space (M, ∆) and mapping h : X → M the Construction 5.1 is a solution to the biometric verification problem.""

    The success event in the problem definition and the output of the verify subroutine are the same predicate, ∆(g(·), enrolled template) ≤ t(λ). The lemma therefore restates the construction's own acceptance rule as a theorem: any system whose verify subroutine returns the indicator of the defining inequality satisfies the definition by construction. No derivation from computational assumptions, no distributional argument, and no proof of Lemma 5.1 is given. The claimed solution is the definition of the subroutine, so the result is true by definition rather than by cryptographic reasoning.

  2. renaming known result [Section 6.1, Definition 6.2, Construction 6.1, Lemma 6.1]
    "Definition 6.2: "Let (M, ∆) define a metric space and g : X → M ... the biometric verification problem is defined by" [followed by the same probability box as Definition 5.1] ... Construction 6.1: "We define a biometric verification system as an interactive protocol ... verifypp(id, x) → b: V returns b := 1{∆(g(x),y)≤t(λ)}" ... Lemma 6.1: "For a metric space (M, ∆) and mapping h : X → M the Construction 6.1 is a solution to the biometric identification problem.""

    Definition 6.2 is not a definition of identification: it reproduces the verification problem verbatim, even retaining the words 'the biometric verification problem'. Construction 6.1 is likewise a near-verbatim copy of Construction 5.1, including the verify subroutine and the phrase 'biometric verification system'. Lemma 6.1 then asserts that this copied verification construction solves the identification problem. The step reduces to renaming the verification result from Section 5 as an identification result; the identification setting described earlier (search over enrolled templates and return ID_i or ⊥) never appears in the formal construction or lemma.

full rationale

The paper's only formal lemmas are self-definitional. Lemma 5.1 is satisfied by construction because Construction 5.1's verify subroutine outputs exactly the indicator of the inequality that Definition 5.1 uses as the success event; no proof or independent argument is supplied. Lemma 6.1 is the same step repeated: Definition 6.2 and Construction 6.1 are verbatim copies of the verification problem and verification construction, so the claimed 'identification' solution is just the verification system relabeled. The abstract's stronger promise, that universal composability and an ideal functionality yield security and privacy properties under computational assumptions, is not derivable from the text because no ideal functionality, simulator, environment, or UC security definition appears anywhere; that absence is a correctness gap rather than a circular step, but it means the advertised framework has no independent formal content beyond the self-referential lemmas. There is no load-bearing self-citation issue in the text. Because the formal results reduce by definition to their own inputs, while part of the failure is absence rather than derivation, a score of 8 is appropriate.

Assumptions & free parameters 3 free parameters · 5 assumptions · 1 invented entities

The paper's central claim depends on an unspecified ideal functionality, an assumed PAC-learnable embedding h, an arbitrary threshold, and an idealized uniform biometric distribution. None of these are derived or supported by data; they are taken as inputs, so the ledger is dominated by unverified assumptions rather than by proven results.

free parameters (3)
  • threshold function t(λ) = unspecified
    Definition 5.1 and Construction 5.1 both use the same arbitrary threshold t(λ) to define success; the paper gives no procedure for setting it, and the 'correctness' of the construction holds for any threshold by construction.
  • database size function ℓ(λ) = unspecified
    The verification and identification problems are parameterized by an arbitrary database size ℓ(λ), with no bound on the biometric distribution or on template distinctness beyond uniqueness with high probability.
  • mapping h in hypothesis class H = unspecified
    The init subroutine assumes h is 'the PAC hypothesis' mapping features to the metric space; the paper provides no trained model, no error bound, and no proof that such an h exists for real biometric distributions.
assumptions (5)
  • standard math Standard metric space axioms (identity of indiscernibles, symmetry, triangle inequality) from Definition 2.1.
    The verification and identification constructions are built on metric spaces and inherit their properties.
  • domain assumption PAC learnability of the embedding h: F → M with low error.
    Construction 5.1's init assumes h is a PAC hypothesis; no distribution over biometric data or error bound is given.
  • domain assumption Biometric source data x are sampled uniformly from X, and enroll/challenge samples are independent.
    Definitions 5.1 and 6.2 sample x_i and x' uniformly from X; real biometric distributions are neither uniform nor independent, so the success probability expression is an idealization.
  • domain assumption Sensors provide quality biometric data; sensor variability is out of scope (Remark 4.1).
    The paper explicitly sets aside the sensor module, which is a load-bearing simplification for any real system.
  • domain assumption Uniqueness of identifiers with high probability.
    Definition 5.1 requires each id_i to be unique with high probability; the construction does not enforce or verify this.
invented entities (1)
  • Ideal functionality for biometric authentication (promised in abstract, not defined)
    purpose: Anchor for a universal composability security and privacy proof; referenced in the abstract and conclusion but never constructed in Sections 2-6.
    Because the functionality is never defined, there is no specification against which an implementation could be checked, and no falsifiable prediction or proof is attached to it.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Framework for the Security and Privacy of Biometric System Constructions under Defined Computational Assumptions." pith.science (2026). https://pith.science/paper/GQU3PRZS

@misc{pith2026241117321,
  author       = {Pith},
  title        = {Pith review of: A Framework for the Security and Privacy of Biometric System Constructions under Defined Computational Assumptions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GQU3PRZS}},
  note         = {Machine review of arXiv:2411.17321}
}
read the original abstract

Biometric systems, while offering convenient authentication, often fall short in providing rigorous security assurances. A primary reason is the ad-hoc design of protocols and components, which hinders the establishment of comprehensive security proofs. This paper introduces a formal framework for constructing secure and privacy-preserving biometric systems. By leveraging the principles of universal composability, we enable the modular analysis and verification of individual system components. This approach allows us to derive strong security and privacy properties for the entire system, grounded in well-defined computational assumptions.

Figures

Figures reproduced from arXiv: 2411.17321 by the authors.

Figure 1
Figure 1. A common biometric system architecture, adapted from work by Jain et al. [30]. 3.1. Biometric System Overview. The canonical explanation of a traditional bio￾metric system is outlined in the seminal work of Jain, Ross, and Prabhakar [29]. A biometric system, at its core, functions as a sophisticated pattern recognition system that processes biological or behavioral characteristics to establish or verify identity. He… view at source ↗
Figure 2
Figure 2. visually illustrates the construction above. We will now look at two concrete examples of biometric verification systems that have been used in practical applications of biometric verification. The first is a verification system using fingerprint features in the Hamming metric, and the second is a verification system that uses face features in the Euclidean metric. U D V DB M w ′ , id h(w ′ ), id b, b ∆( ′ b, b ′ ) … view at source ↗
Figure 3
Figure 3. A biometric identification system model with indications of the generic information flows based on work by Jain et al. [30] and Con￾struction 6.1. In the above figure, we assume that the init method of Construction 6.1 has already been run. Lemma 6.1. For a metric space (M, ∆) and mapping h : X → M the Construction 6.1 is a solution to the biometric identification problem. Example 6.1 (Face Verification in the Eucli… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 64 canonical work pages

  1. [1]

    The 2022 duo trusted access report

    D. Lewis, “The 2022 duo trusted access report.” [Online]. Available: https://duo.com/resources/ ebooks/the-2022-duo-trusted-access-report

    work page 2022

  2. [2]

    Facing the future: the impact of apple faceid,

    A. Bud, “Facing the future: the impact of apple faceid,” vol. 2018, no. 1, pp. 5–7. 18 S. GRIESON ET AL

    work page 2018

  3. [4]

    A. Support. About face id advanced techonology. Apple. [Online]. Available: https: //support.apple.com/en-gb/HT208108

    work page

  4. [5]

    Biometric authentication on iphone and android: Usability, perceptions, and influences on adoption,

    R. Bhagavatula, B. Ur, K. Iacovino, S. M. Kywe, L. F. Cranor, and M. Savvides, “Biometric authentication on iphone and android: Usability, perceptions, and influences on adoption,” in USEC ’15: Workshop on Usable Security, 8 February 2015, San Diego, CA: Proceedings. Internet Society, pp. 1–10

    work page 2015

  5. [6]

    Big other: surveillance capitalism and the prospects of an information civilization,

    S. Zuboff, “Big other: surveillance capitalism and the prospects of an information civilization,” in Journal of Information Technology , vol. 30, pp. 75–89

    work page

  6. [7]

    ——, The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power: Barack Obama’s Books of 2019 . Profile

    work page 2019

  7. [8]

    Open sesame! user perception of physical, biometric, and behavioural authentication concepts to open doors,

    L. Mecke, K. Pfeuffer, S. Prange, and F. Alt, “Open sesame! user perception of physical, biometric, and behavioural authentication concepts to open doors,” in Proceedings of the 17th International Conference on Mobile and Ubiquitous Multimedia , ser. MUM 2018. Association for Computing Machinery, pp. 153–159

    work page 2018

  8. [9]

    Privacy concerns and disclosure of biometric and behav- ioral data for travel,

    A. Ioannou, I. Tussyadiah, and Y. Lu, “Privacy concerns and disclosure of biometric and behav- ioral data for travel,” vol. 54, p. 102122

    work page

Show all 64 references
  1. [10]

    European Commission

    General data protection regulation (gdpr). European Commission. [Online]. Available: https://gdpr-info.eu/

  2. [11]

    California State legislature

    California consumer privacy act of 2018. California State legislature. [Online]. Available: https://leginfo.legislature.ca.gov/faces/codes displayText.xhtml?division=3.&part= 4.&lawCode=CIV&title=1.81.5

    2018

  3. [12]

    Universally composable security: a new paradigm for cryptographic protocols,

    R. Canetti, “Universally composable security: a new paradigm for cryptographic protocols,” in Proceedings 42nd IEEE Symposium on Foundations of Computer Science , pp. 136–145

  4. [13]

    Fuzzy extractors: How to generate strong keys from biomet- rics and other noisy data,

    Y. Dodis, L. Reyzin, and A. Smith, “Fuzzy extractors: How to generate strong keys from biomet- rics and other noisy data,” in Advances in Cryptology - EUROCRYPT 2004 , C. Cachin and J. L. Camenisch, Eds. Springer Berlin Heidelberg, pp. 523–540

    2004

  5. [14]

    Robust fuzzy extractors and authenticated key agree- ment from close secrets,

    Y. Dodis, J. Katz, L. Reyzin, and A. Smith, “Robust fuzzy extractors and authenticated key agree- ment from close secrets,” in Advances in Cryptology - CRYPTO 2006 , C. Dwork, Ed. Springer Berlin Heidelberg, pp. 232–250

    2006

  6. [15]

    Secure remote authentication using biometric data,

    X. Boyen, Y. Dodis, J. Katz, R. Ostrovsky, and A. Smith, “Secure remote authentication using biometric data,” in Advances in Cryptology – EUROCRYPT 2005 , R. Cramer, Ed. Springer Berlin Heidelberg, pp. 147–163

    2005

  7. [16]

    A logical calculus of the ideas immanent in nervous activity,

    W. S. McCulloch and W. Pitts, “A logical calculus of the ideas immanent in nervous activity,” The bulletin of mathematical biophysics , vol. 5, no. 4, pp. 115–133, 12 1943

    1943

  8. [17]

    A Fast Learning Algorithm for Deep Belief Nets,

    G. E. Hinton, S. Osindero, and Y.-W. Teh, “A Fast Learning Algorithm for Deep Belief Nets,” Neural Computation, vol. 18, no. 7, pp. 1527–1554, 07 2006

    2006

  9. [18]

    Minsky and S

    M. Minsky and S. A. Papert, Perceptrons; An Introduction to Computational Geometry . The MIT Press, 1969

    1969

  10. [19]

    Heuristic self-organization in problems of engineering cybernetics,

    A. G. Ivakhnenko, “Heuristic self-organization in problems of engineering cybernetics,” Automat- ica, vol. 6, no. 2, pp. 207–219, 1970

    1970

  11. [20]

    Polynomial theory of complex systems,

    ——, “Polynomial theory of complex systems,” IEEE Transactions on Systems, Man, and Cyber- netics, vol. SMC-1, no. 4, pp. 364–378, 1971. BIOMETRIC SYSTEM CONSTRUCTIONS 19

    1971

  12. [21]

    A theory of adaptive pattern classifiers,

    S. Amari, “A theory of adaptive pattern classifiers,” IEEE Transactions on Electronic Computers, vol. EC-16, no. 3, pp. 299–307, 1967

    1967

  13. [22]

    Delving deep into rectifiers: Surpassing human-level performance on imagenet classification,

    K. He, X. Zhang, S. Ren, and J. Sun, “Delving deep into rectifiers: Surpassing human-level performance on imagenet classification,” in 2015 IEEE International Conference on Computer Vision (ICCV) , 2015, pp. 1026–1034

    2015

  14. [23]

    Visual feature extraction by a multilayered network of analog threshold elements,

    K. Fukushima, “Visual feature extraction by a multilayered network of analog threshold elements,” IEEE Transactions on Systems Science and Cybernetics , vol. 5, no. 4, pp. 322–333, 1969

    1969

  15. [24]

    Neocognitron: A self-organizing neural network model for a mech- anism of visual pattern recognition,

    K. Fukushima and S. Miyake, “Neocognitron: A self-organizing neural network model for a mech- anism of visual pattern recognition,” in Competition and Cooperation in Neural Nets , S.-i. Amari and M. A. Arbib, Eds. Berlin, Heidelberg: Springer Berlin Heidelberg, 1982, pp. 267–285

    1982

  16. [25]

    Imagenet classification with deep convolutional neural networks,

    A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classification with deep convolutional neural networks,” in Advances in Neural Information Processing Systems , F. Pereira, C. Burges, L. Bottou, and K. Weinberger, Eds., vol. 25. Curran Associates, Inc., 2012

    2012

  17. [26]

    Backpropagation applied to handwritten zip code recognition,

    Y. LeCun, B. Boser, J. S. Denker, D. Henderson, R. E. Howard, W. Hubbard, and L. D. Jackel, “Backpropagation applied to handwritten zip code recognition,”Neural Computation, vol. 1, no. 4, pp. 541–551, 12 1989

    1989

  18. [27]

    Gradient-based learning applied to document recognition,

    Y. Lecun, L. Bottou, Y. Bengio, and P. Haffner, “Gradient-based learning applied to document recognition,” Proceedings of the IEEE, vol. 86, no. 11, pp. 2278–2324, 1998

    1998

  19. [28]

    A neural network for speaker- independent isolated word recognition,

    K. Yamaguchi, K. Sakamoto, T. Akabane, and Y. Fujimoto, “A neural network for speaker- independent isolated word recognition,” in First International Conference on Spoken Language Processing (ICSLP 1990), 1990, pp. 1077–1080

    1990

  20. [29]

    An introduction to biometric recognition,

    A. K. Jain, A. Ross, and S. Prabhakar, “An introduction to biometric recognition,” vol. 14, no. 1, pp. 4–20

  21. [30]

    A. K. Jain, P. Flynn, and A. Ross, Handbook of biometrics. Springer Science & Business Media

  22. [31]

    “if it wasn’t secure, they would not use it in the movies

    V. Zimmermann and N. Gerber, ““if it wasn’t secure, they would not use it in the movies” – security perceptions and user acceptance of authentication technologies,” in Human Aspects of Information Security, Privacy and Trust , T. Tryfonas, Ed. Springer International Publishing...

  23. [32]

    Acceptance of biometric authentication security technology on mobile devices,

    W. Ratjeana Malatji, R. van Eck, and T. Zuva, “Acceptance of biometric authentication security technology on mobile devices,” in 2020 2nd International Multidisciplinary Information Technol- ogy and Engineering Conference (IMITEC) , pp. 1–5

    2020

  24. [33]

    S. Z. Li and A. K. Jain, Eds., Handbook of Face Recognition, 2nd ed. Springer London

  25. [34]

    Security and accuracy of fingerprint-based biometrics: A review,

    W. Yang, S. Wang, J. Hu, G. Zheng, and C. Valli, “Security and accuracy of fingerprint-based biometrics: A review,” vol. 11, no. 2

  26. [35]

    K. W. Bowyer, K. P. Hollingsworth, and P. J. Flynn, A Survey of Iris Biometrics Research: 2008–2010. Springer London, pp. 23–61

    2008

  27. [36]

    A comprehensive review on iris image-based biometric system,

    J. Winston and D. J. Hemanth, “A comprehensive review on iris image-based biometric system,” vol. 23, no. 19, pp. 9361–9384

  28. [37]

    Distinctive image features from scale-invariant keypoints,

    D. G. Lowe, “Distinctive image features from scale-invariant keypoints,” vol. 60, no. 2, pp. 91–110

  29. [38]

    Histograms of oriented gradients for human detection,

    N. Dalal and B. Triggs, “Histograms of oriented gradients for human detection,” in 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05) , vol. 1, pp. 886–893 vol. 1

    2005

  30. [39]

    Feature-domain super-resolution framework for gabor-based face and iris recognition,

    K. Nguyen, S. Sridharan, S. Denman, and C. Fookes, “Feature-domain super-resolution framework for gabor-based face and iris recognition,” in 2012 IEEE Conference on Computer Vision and Pattern Recognition, pp. 2642–2649. 20 S. GRIESON ET AL

    2012

  31. [40]

    Face recognition using holistic fourier invariant features,

    J. H. Lai, P. C. Yuen, and G. C. Feng, “Face recognition using holistic fourier invariant features,” vol. 34, no. 1, pp. 95–109

  32. [41]

    An efficient fingerprint verification system using integrated wavelet and fourier–mellin invariant transform,

    A. T. B. Jin, D. N. C. Ling, and O. T. Song, “An efficient fingerprint verification system using integrated wavelet and fourier–mellin invariant transform,” vol. 22, no. 6, pp. 503–513

  33. [42]

    Eigenfaces for recognition,

    M. Turk and A. Pentland, “Eigenfaces for recognition,” vol. 3, no. 1, pp. 71–86

  34. [43]

    Deep learning for biometrics: A survey,

    K. Sundararajan and D. L. Woodard, “Deep learning for biometrics: A survey,” vol. 51, no. 3

  35. [44]

    Biometrics recognition using deep learning: a survey,

    S. Minaee, A. Abdolrashidi, H. Su, M. Bennamoun, and D. Zhang, “Biometrics recognition using deep learning: a survey,” vol. 56, no. 8, pp. 8647–8695

  36. [45]

    Fingernet: An unified deep network for fingerprint minutiae extraction,

    Y. Tang, F. Gao, J. Feng, and Y. Liu, “Fingernet: An unified deep network for fingerprint minutiae extraction,” in 2017 IEEE International Joint Conference on Biometrics (IJCB) , pp. 108–116

    2017

  37. [46]

    Arcface: Additive angular margin loss for deep face recognition,

    J. Deng, J. Guo, J. Yang, N. Xue, I. Kotsia, and S. Zafeiriou, “Arcface: Additive angular margin loss for deep face recognition,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 44, no. 10, pp. 5962–5979, 2022

    2022

  38. [47]

    A survey of emerging biometric modalities,

    S. Chauhan, A. Arora, and A. Kaul, “A survey of emerging biometric modalities,” vol. 2, pp. 213–218, proceedings of the International Conference and Exhibition on Biometrics Technology

  39. [48]

    Learning discriminative binary codes for finger vein recognition,

    X. Xi, L. Yang, and Y. Yin, “Learning discriminative binary codes for finger vein recognition,” vol. 66, pp. 26–33

  40. [49]

    FaceNet: A unified embedding for face recognition and clustering,

    F. Schroff, D. Kalenichenko, and J. Philbin, “FaceNet: A unified embedding for face recognition and clustering,” in 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) . IEEE

    2015

  41. [50]

    Altered fingerprints: Analysis and detection,

    S. Yoon, J. Feng, and A. K. Jain, “Altered fingerprints: Analysis and detection,” vol. 34, no. 3, pp. 451–464

  42. [51]

    Biometric template security,

    A. K. Jain, K. Nandakumar, and A. Nagar, “Biometric template security,” vol. 2008

    2008

  43. [52]

    Biometric systems: Privacy and secrecy aspects,

    T. Ignatenko and F. M. J. Willems, “Biometric systems: Privacy and secrecy aspects,” vol. 4, no. 4, pp. 956–973

  44. [53]

    A theoretical analysis of authentication, privacy, and reusability across secure biometric systems,

    Y. Wang, S. Rane, S. C. Draper, and P. Ishwar, “A theoretical analysis of authentication, privacy, and reusability across secure biometric systems,” vol. 7, no. 6, pp. 1825–1840

  45. [54]

    Fuzzy extractors for biometric identification,

    N. Li, F. Guo, Y. Mu, W. Susilo, and S. Nepal, “Fuzzy extractors for biometric identification,” in 2017 IEEE 37th International Conference on Distributed Computing Systems (ICDCS) , pp. 667–677

    2017

  46. [55]

    Highly robust biometric smart card design,

    A. Noore, “Highly robust biometric smart card design,” vol. 46, no. 4, pp. 1059–1063

  47. [56]

    Provable security analysis of fido2,

    M. Barbosa, A. Boldyreva, S. Chen, and B. Warinschi, “Provable security analysis of fido2,” in Advances in Cryptology – CRYPTO 2021, T. Malkin and C. Peikert, Eds. Springer International Publishing, pp. 125–156

    2021

  48. [57]

    A. Jain, R. Bolle, and S. Pankanti, Biometrics: Personal Identification in Networked Society , A. K. Jain, R. Bolle, and S. Pankanti, Eds. Springer US

  49. [58]

    An introduction evaluating bio- metric systems,

    P. Phillips, A. F. Martin, C. L. Wilson, and M. A. Przybocki, “An introduction evaluating bio- metric systems,” vol. 33, no. 2, pp. 56–63

  50. [59]

    Biometric recognition: security and privacy concerns,

    S. Prabhakar, S. Pankanti, and A. K. Jain, “Biometric recognition: security and privacy concerns,” vol. 1, no. 2, pp. 33–42

  51. [60]

    R. M. Bolle, J. H. Connell, S. Pankanti, N. K. Ratha, and A. W. Senior, Authentication and Biometrics. Springer New York, pp. 17–30

  52. [61]

    Biometrics: a tool for information security,

    A. K. Jain, A. Ross, and S. Pankanti, “Biometrics: a tool for information security,” vol. 1, no. 2, pp. 125–143. BIOMETRIC SYSTEM CONSTRUCTIONS 21

  53. [62]

    Fast computation of the performance evaluation of biometric systems: Application to multibiometrics,

    R. Giot, M. El-Abed, and C. Rosenberger, “Fast computation of the performance evaluation of biometric systems: Application to multibiometrics,” vol. 29, no. 3, pp. 788–799, special Section: Recent Developments in High Performance Computing and Security

  54. [63]

    A. Support. About touch id advanced security techonology. Apple. [Online]. Available: https://support.apple.com/en-gb/105095

  55. [64]

    G. Store. Pixel 8 pro tech specs. Google. [Online]. Available: https://store.google.com/product/ pixel 8 pro specs?hl=en-GB

  56. [65]

    Fundamentals of biometric authentication technologies,

    J. L. Wayman, “Fundamentals of biometric authentication technologies,” vol. 01, no. 01, pp. 93–113. Sam Grierson, Edinburgh Napier University, Edinburgh, UK Email address : s.grierson2@napier.ac.uk William J Buchanan, Blockpass ID Lab, Edinburgh Napier University, Edinburgh, U...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.