Next Article in Journal
A Novel Deep Learning Model for Motor Imagery Classification in Brain–Computer Interfaces
Previous Article in Journal
A KeyBERT-Enhanced Pipeline for Electronic Information Curriculum Knowledge Graphs: Design, Evaluation, and Ontology Alignment
Previous Article in Special Issue
Real-Time Object Detection Model for Electric Power Operation Violation Identification
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Information
Volume 16
Issue 7
10.3390/info16070581
information-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Systematic Review

Advancing Secure Face Recognition Payment Systems: A Systematic Literature Review

by
M. Haswin Anugrah Pratama
1,2,*,
Achmad Rizal
2 and
Indrarini Dyah Irawati
3
1
Artificial Intelligence Department, Bank Rakyat Indonesia, Jakarta 10210, Indonesia
2
School of Electrical Engineering, Telkom University, Bandung 40257, Indonesia
3
School of Applied Science, Telkom University, Bandung 40257, Indonesia
*
Author to whom correspondence should be addressed.
Information 2025, 16(7), 581; https://doi.org/10.3390/info16070581
Submission received: 27 May 2025 / Revised: 27 June 2025 / Accepted: 30 June 2025 / Published: 7 July 2025
(This article belongs to the Special Issue Computer Vision for Security Applications, 2nd Edition)
Browse Figures
Versions Notes

Abstract

In the digital era, face recognition technology has emerged as a promising solution for enhancing payment system security and convenience. This systematic literature review examines face recognition advancements in payment security following the PRISMA framework. From 219 initially identified articles, we selected 10 studies meeting our technical criteria. The findings reveal significant progress in deep learning approaches, multimodal feature integration, and transformer-based architectures. Current trends emphasize multimodal systems combining RGB with IR and depth data for sophisticated attack detection. Critical challenges remain in cross-dataset generalization, evaluation standardization, computational efficiency, and combating advanced threats including deepfakes. This review identifies technical limitations and provides direction for developing robust facial recognition technologies for widespread payment adoption.

1. Introduction

In the digital era, various technologies are rapidly advancing, including technologies for transactions and payments. One of these is the cashless or non-cash payment method, which is a method of payment for certain transactions without using physical cash. In Indonesia, non-cash payment instruments are divided into three types: paper-based, card-based, and electronic types [1]. Unfortunately, all three of these instruments still require physical devices that need to be carried around. In comparison, China has implemented biometric-based payments, one of which is face recognition payment (FRP) [2]. FRP technology allows people to make payments using their faces, which are linked to their bank accounts [3]. In 2018, it was reported that FRP had been used by 61 million registered users in China, and this number was estimated to reach 760 million by 2022 [4].
The development of FRP is driven by advancements in facial recognition methods that continue to evolve. However, facial recognition methods still face challenges in terms of accuracy and security. Accuracy challenges are influenced by environmental conditions such as lighting, pose, expression, aging, occlusion, low resolution, facial scars, and plastic surgery [5,6,7,8,9]. Meanwhile, security challenges are influenced by various types of attacks, such as print attacks, replay attacks, mask attacks, partial occlusion attacks, makeup attacks, deepfake attacks, statue attacks, and plastic surgery attacks [10,11,12,13,14,15].
These challenges must be addressed to implement a robust and general FRP system capable of adapting to various existing scenarios. Therefore, a deeper understanding of the various challenges and solutions involved in the development of FRP systems is required. As an initial step, a literature review on FRP is essential to explore the various aspects of this technology. By conducting a literature review, we can learn about the approaches that have been previously applied, particularly in addressing issues related to accuracy and security. Various methods and algorithms in facial recognition that have been used in FRP systems will be evaluated, including their strengths and weaknesses. Additionally, this literature review will help us to understand the various attack threats that can compromise the integrity of the system, enabling the design of more robust, reliable, and adaptable FRP systems that can handle different scenarios.

2. Methods

This study employed a systematic literature review methodology following PRISMA guidelines [16] to ensure transparent and consistent study selection. The methodology was designed to comprehensively examine algorithmic advancements in face recognition payment systems while maintaining rigorous selection criteria.

2.1. Research Question

This systematic review addressed three primary research questions:
  • RQ1: What are the key technological advancements in face recognition methods that have improved payment system security over the past five years?
  • RQ2: How effective are current face anti-spoofing techniques against emerging threats to face recognition payment systems?
  • RQ3: What are the critical challenges involved in deploying face recognition payment systems in real-world environments?

2.2. Search Strategy

The literature search was conducted using three major databases, Scopus, IEEE Xplore, and Springer Link, selected for their reputation for providing high-quality scientific articles on information technology and facial recognition systems. The search was limited to publications published from 1 January 2019 to 31 December 2024 to ensure current relevance. A consistent set of keywords was used, “face payment” OR “facial payment” OR “face recognition payment” OR “facial recognition payment”, to capture various relevant terms in the literature.

2.3. PRISMA Process

The PRISMA process’ implementation [16] is illustrated in Figure 1, providing a systematic workflow from identification through to inclusion stages (Supplementary Materials). The identification stage began with a literature search across three primary databases, Scopus, Springer Link, and IEEE Xplore, yielding 219 articles (Scopus: 82; Springer Link: 123; IEEE Xplore: 14), as shown in Figure 2. After removing duplicates (21), non-articles (38), and non-English publications (4), 146 articles remained for screening.
During screening, all 146 articles were reviewed for alignment with the inclusion criteria. Due to access limitations, 87 articles could not be retrieved, leaving 59 articles applicable for eligibility assessment. In the eligibility stage, articles were evaluated using technical criteria focusing on algorithmic advancements and computational methods. Based on this evaluation, 35 articles were excluded for not meeting technical requirements, with 23 articles deemed suitable for inclusion.
In the final inclusion stage, 10 articles were selected based on relevance, quality, and making a significant contribution to the research topic, as listed in Table 1. These articles provide a strong foundation for understanding current developments in face recognition payment systems.

3. Results

The selected articles provide comprehensive insights into the most significant technological advancements, anti-spoofing effectiveness, and critical deployment challenges related to face recognition payment systems, as detailed in Table 1. These findings directly address our three research questions and highlight the key developments that are shaping the future of secure payment technologies.

3.1. Technological Advancements in Face Recognition Methods (RQ1)

The most significant advancement in face recognition for payment systems centers on three breakthrough areas that have fundamentally improved system performance. Deep learning-based approaches represent the primary innovation driver, with the most notable achievements including Xiao et al. [17] developing a MobileFaceNet-based [27] anti-spoofing algorithm incorporating Coordinate Attention (CA) [28] achieving 1.39% ACER with a 45 ms processing time, outperforming methods like CDCN and CDCN++ [29]; Guo et al. [18] introducing MAHA-Net for deepfake detection combining Multiscale Attention and Halo Attention to achieve 97.12% accuracy on high-quality FaceForensics++ (FF++) [30] data; and Lu et al. [24] proposing a Heterogeneous Kernel–Convolutional Neural Network (HK-CNN) using heterogeneous kernels, achieving 99.82% accuracy on NUAA [31] and 99.17% accuracy on CASIA-FASD [32] while reducing the training time by 19.4%. These innovations demonstrate the critical balance between performance and efficiency necessary for practical deployment.
Multimodal feature integration represents the second major advancement area, with exceptional performance demonstrated by leading studies. Yu et al. [19] combined Vision Transformer (ViT) [33] with Masked Autoencoder (MAE) [34] for multimodal face anti-spoofing, incorporating Adaptive Multimodal Adapter (AMA) and Modality-Asymmetric Masked Autoencoder (M2A2E), achieving cross-attack performance of 6.08% ACER on WMCA [35], 2.12% ACER on CASIA-SURF [36], and 8.36% ACER on CeFA [37]. Li et al. [22] developed a Middle-Shallow Feature Aggregation (MSFA) model integrating RGB, Depth, and IR data, achieving a remarkable 0.079% ACER on CASIA-SURF [36], representing near-perfect performance on standard benchmarks. Parkin et al. [25] demonstrated superior multimodal performance with a 99.87% TPR using a ResNet-based [38] architecture with a Squeeze-and-Excitation (SE) [39] module and Multi-Level Feature Aggregation (MLFA) combining RGB, IR, and Depth data, consistently outperforming single-modality approaches.
Transformer-based architectures constitute the third breakthrough area, with Wang et al. [20] introducing Dynamic Feature Queue (DFQ) and Progressive Training Strategy (PTS), achieving a 4.73% ACER and an AUC of 98.38% on a challenging SuHiFiMask [40] dataset through innovative multi-class reformulation of binary classification problems, demonstrating enhanced generalization capabilities against unseen attacks.

3.2. Effectiveness of Current Face Anti-Spoofing Techniques (RQ2)

Anti-spoofing technique effectiveness varies significantly across attack types and deployment conditions, revealing critical performance patterns. Against conventional attacks, Lin et al. [26] developed a face liveness detection method combining Remote Photoplethysmography (rPPG) [41] features and a Contextual Patch-Based CNN (CP-CNN), achieving a perfect performance with a 0.0% Equal Error Rate on MSU-MFSD [42], 3DMAD [43], and HKBU-Mars V1 [44] datasets for print/replay and 3D mask attacks, respectively, and achieving a 1.8% EER on CASIA-FASD [32] and a 15.1% ACER on OULU-NPU [45], demonstrating that traditional threats can be effectively countered under controlled conditions. Lu et al. [24] achieved impressive results with their HK-CNN model, demonstrating a 99.82% accuracy on NUAA [31] and a 99.17% accuracy on CASIA-FASD [32], with a significant reduction in the Half Total Error Rate (HTER) to 1.03, compared to 2.27 for DB-CNN [46] and 14.26 for VCST [47].
However, the most critical finding is the significant performance gap when facing advanced threats and real-world scenarios. Deepfake detection shows consistent degradation, with Guo et al. [18] achieving a 97.12% accuracy on high-quality FaceForensics++ data but dropping to 91.26% on low-quality compressed images, while Zou et al. [21] developed a DIFLD framework demonstrating similar challenges, with a 96.56% accuracy on medium-quality FaceForensics++ (FF++) [30], decreasing to 90.29% on low-quality datasets. This pattern indicates fundamental vulnerability to compression and environmental factors that are common in practical deployment scenarios.
Cross-dataset generalization represents the most severe limitation, with Wang et al. [23] developing a face anti-spoofing model using Siamese network [48] architecture, achieving a 4.17% ACER against unknown presentation attacks from a SiW [49] dataset but demonstrating substantial performance degradation with a 23.9% HTER on CASIA-FASD [32] and a 38.0% HTER on Replay-Attack [50] during cross-dataset testing. This issue represents a fundamental barrier for practical deployment across diverse populations and environments not represented in training data.

3.3. Critical Challenges and Limitations in Real-World Deployment (RQ3)

The most significant deployment challenges center on three critical areas that limit practical implementation. Environmental robustness remains the primary concern, with multiple studies including Xiao et al. [17], Guo et al. [18], and Wang et al. [20] identifying performance degradation under variable lighting, non-frontal poses, and image quality variations. These environmental factors significantly impact system reliability in real-world payment scenarios where controlled conditions cannot be guaranteed.
Computational efficiency presents the second major challenge, particularly evident in the accuracy–speed trade-off that constrains deployment on resource-limited payment devices. While advanced methods like that of Yu et al. [19] achieve superior detection capabilities with Vision Transformer architectures, they require computational resources exceeding typical payment hardware specifications. Xiao et al. [17] demonstrated progress, with their MobileFaceNet-based approach achieving a 45ms processing time for edge devices, while Li et al. [22] focused on reducing computational complexity with their MSFA model compared to previous approaches like MLFA [25]. This fundamental tension between algorithmic sophistication and practical deployment constraints necessitates continued optimization research.
The third critical challenge involves dataset and evaluation limitations, including over-reliance on potentially outdated datasets like CASIA-FASD [32] and Replay-Attack [50] that may not reflect current attack sophistication or demographic diversity. The lack of standardized metrics across studies, with different research utilizing various measures including ACER, AUC, EER, TPR, and HTER, as shown in Table 1, hampers systematic algorithm comparison and impedes progress toward payment-specific evaluation frameworks that adequately represent real-world deployment scenarios including transaction-level security demands and edge deployment limitations typical of commercial payment environments.

4. Discussion

This systematic literature review focused specifically on technical and algorithmic contributions to face recognition payment systems, revealing critical insights about current capabilities and fundamental limitations. The deliberate technical scope enabled comprehensive analysis of computational methods and algorithmic innovations while necessarily excluding broader deployment considerations. Our methodology prioritized studies with concrete technical contributions and quantitative performance metrics, ensuring that its rigorous analysis aligned with our research objectives.
The most significant finding concerns the substantial performance gap between laboratory validation and real-world deployment scenarios. While the reviewed algorithms demonstrate excellent performance on controlled datasets, they show consistent degradation when tested across different datasets or environmental conditions. This laboratory-to-real-world gap represents the primary barrier for practical payment system deployment, as these systems must function reliably across diverse populations, environmental conditions, and hardware configurations not represented in training data.
Computational efficiency emerges as the second critical limitation, particularly the fundamental tension between algorithmic sophistication and practical deployment constraints. Advanced transformer-based architectures and multimodal approaches achieve superior detection performance but require computational resources exceeding typical payment hardware capabilities. This constraint necessitates future research prioritizing efficiency optimization alongside accuracy improvements, potentially through hardware–software co-design approaches or specialized lightweight architectures for payment applications.
The rapidly evolving threat landscape, particularly sophisticated attacks like deepfakes and adaptive presentation attacks, represents an ongoing challenge requiring continuous algorithmic advancement. Current anti-spoofing techniques show consistent vulnerability to image quality degradation and novel attack combinations, indicating the need for more robust detection algorithms capable of handling previously unseen attack vectors and maintaining performance across varying input conditions.

5. Conclusions

This systematic literature review examined algorithmic advancements in face recognition payment systems using PRISMA methodology, analyzing 10 key studies selected from 219 initially identified articles. The findings reveal significant progress in deep learning approaches, multimodal feature integration, and transformer-based architectures that have substantially improved computational security against conventional spoofing attacks. However, critical technical gaps persist including cross-dataset generalization difficulties, inconsistent evaluation metrics, computational efficiency trade-offs, and limited effectiveness against advanced threats such as deepfakes and adaptive attacks.
Future research should prioritize developing robust cross-dataset generalization techniques to address performance degradation when models encounter new datasets or environmental conditions. Establishing standardized evaluation frameworks with unified metrics will enable meaningful algorithmic performance comparison and systematic advancement in the field. The development of payment-specific benchmark datasets reflecting real-world deployment scenarios is essential for practical algorithm validation. Additionally, advancing adaptive threat detection capabilities to address evolving attack strategies and optimizing computational efficiency to bridge the gap between high-performance algorithms and resource-limited payment devices represent critical research priorities.
By building on the technical foundations identified in this review, researchers can develop facial recognition technologies that maintain robust security through algorithmic innovation while offering the computational efficiency needed for widespread deployment across various payment applications. The systematic analysis conducted here establishes a foundation for future algorithmic research to advance secure, efficient, and practical face recognition payment systems.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/info16070581/s1, PRISMA 2020 Checklist. Reference [51] is citied in the Supplementary Materials.

Funding

This research received no external funding.

Conflicts of Interest

M. Haswin Anugrah Pratama is employed by Bank Rakyat Indonesia. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

References

  1. Bank Indonesia. Indonesia Payment Systems Blueprint 2025: Navigating the National Payment Systems in the Digital Era. Bank Indonesia. 2019, pp. 1–50. Available online: https://www.bi.go.id/en/publikasi/kajian/Documents/Indonesia-Payment-Systems-Blueprint-2025.pdf (accessed on 15 May 2025).
  2. Zhang, W.K.; Kang, M.J. Factors Affecting the Use of Facial-Recognition Payment: An Example of Chinese Consumers. IEEE Access 2019, 7, 154360–154374. [Google Scholar] [CrossRef]
  3. Li, C.; Li, H. Disentangling facial recognition payment service usage behavior: A trust perspective. Telemat. Inf. 2023, 77, 101939. [Google Scholar] [CrossRef]
  4. Liu, Y.; Yan, W.; Hu, B. Resistance to facial recognition payment in China: The influence of privacy-related factors. Telecomm. Policy 2021, 45, 102155. [Google Scholar] [CrossRef]
  5. Gururaj, H.L.; Soundarya, B.C.; Priya, S.; Shreyas, J.; Flammini, F. A Comprehensive Review of Face Recognition Techniques, Trends, and Challenges. IEEE Access 2024, 12, 107903–107926. [Google Scholar] [CrossRef]
  6. Jayaraman, U.; Gupta, P.; Gupta, S.; Arora, G.; Tiwari, K. Recent development in face recognition. Neurocomputing 2020, 408, 231–245. [Google Scholar] [CrossRef]
  7. Oloyede, M.O.; Hancke, G.P.; Myburgh, H.C. A review on face recognition systems: Recent approaches and challenges. Multimed. Tools Appl. 2020, 79, 27891–27922. [Google Scholar] [CrossRef]
  8. Singh, S.; Prasad, S.V.A.V. Techniques and Challenges of Face Recognition: A Critical Review. Procedia Comput. Sci. 2018, 143, 536–543. [Google Scholar] [CrossRef]
  9. Adjabi, I.; Ouahabi, A.; Benzaoui, A.; Taleb-Ahmed, A. Past, Present, and Future of Face Recognition: A Review. Electronics 2020, 9, 1188. [Google Scholar] [CrossRef]
  10. Abdullakutty, F.; Elyan, E.; Johnston, P. A review of state-of-the-art in Face Presentation Attack Detection: From early development to advanced deep learning and multi-modal fusion methods. Inf. Fusion 2021, 75, 55–69. [Google Scholar] [CrossRef]
  11. Khairnar, S.; Gite, S.; Kotecha, K.; Thepade, S.D. Face Liveness Detection Using Artificial Intelligence Techniques: A Systematic Literature Review and Future Directions. Big Data Cogn. Comput. 2023, 7, 37. [Google Scholar] [CrossRef]
  12. Ming, Z.; Visani, M.; Luqman, M.M.; Burie, J.-C. A Survey on Anti-Spoofing Methods for Facial Recognition with RGB Cameras of Generic Consumer Devices. J. Imaging 2020, 6, 139. [Google Scholar] [CrossRef] [PubMed]
  13. Pooshideh, M.; Beheshti, A.; Qi, Y.; Farhood, H.; Simpson, M.; Gatland, N.; Soltany, M. Presentation Attack Detection: A Systematic Literature Review. ACM Comput. Surv. 2025, 57, 1–32. [Google Scholar] [CrossRef]
  14. Sharma, D.; Selwal, A. A survey on face presentation attack detection mechanisms: Hitherto and future perspectives. Multimed. Syst. 2023, 29, 1527–1577. [Google Scholar] [CrossRef]
  15. Yu, Z.; Qin, Y.; Li, X.; Zhao, C.; Lei, Z.; Zhao, G. Deep Learning for Face Anti-Spoofing: A Survey. IEEE Trans. Pattern Anal. Mach. Intell. 2022, 45, 5609–5631. [Google Scholar] [CrossRef]
  16. Moher, D.; Liberati, A.; Tetzlaff, J.; Altman, D.G. Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med. 2009, 6, e1000097. [Google Scholar] [CrossRef]
  17. Xiao, J.; Wang, W.; Zhang, L.; Liu, H. A MobileFaceNet-Based Face Anti-Spoofing Algorithm for Low-Quality Images. Electronics 2024, 13, 2801. [Google Scholar] [CrossRef]
  18. Guo, S.; Li, Q.; Gao, M.; Zhang, G.; Pan, J.; Jeon, G. Deep Learning-Based Face Forgery Detection for Facial Payment Systems. IEEE Consum. Electron. Mag. 2024, 14, 80–86. [Google Scholar] [CrossRef]
  19. Yu, Z.; Cai, R.; Cui, Y.; Liu, X.; Hu, Y.; Kot, A.C. Rethinking Vision Transformer and Masked Autoencoder in Multimodal Face Anti-Spoofing. Int. J. Comput. Vis. 2024, 132, 5217–5238. [Google Scholar] [CrossRef]
  20. Wang, K.; Huang, M.; Zhang, G.; Yue, H.; Zhang, G.; Qiao, Y. Dynamic Feature Queue for Surveillance Face Anti-spoofing via Progressive Training. In Proceedings of the 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Vancouver, BC, Canada, 17–24 June 2023; pp. 6372–6379. [Google Scholar]
  21. Zou, Y.; Luo, C.; Zhang, J. DIFLD: Domain invariant feature learning to detect low-quality compressed face forgery images. Complex Intell. Syst. 2024, 10, 357–368. [Google Scholar] [CrossRef]
  22. Li, C.; Li, Z.; Sun, J.; Li, R. Middle-shallow feature aggregation in multimodality for face anti-spoofing. Sci. Rep. 2023, 13, 9870. [Google Scholar] [CrossRef]
  23. Wang, Y.; Pei, M.; Nie, Z.; Qi, X. Face Anti-spoofing Based on Client Identity Information and Depth Map. In Image and Graphics, Proceedings of the ICIG 2023, Nanjing, China, 22–24 September 2023; Lu, H., Ouyang, W., Huang, H., Lu, J., Liu, R., Dong, J., Xu, M., Eds.; Lecture Notes in Computer Science; Springer: Berlin/Heidelberg, Germany, 2023; Volume 14355. [Google Scholar]
  24. Lu, X.; Tian, Y. Heterogeneous Kernel Based Convolutional Neural Network for Face Liveness Detection. In Bio-Inspired Computing: Theories and Applications, Proceedings of the BIC-TA 2019, Zhengzhou, China, 22–25 November 2019; Communications in Computer and Information Science; Springer: Singapore, 2020; Volume 1160. [Google Scholar]
  25. Parkin, A.; Grinchuk, O. Recognizing Multi-Modal Face Spoofing with Face Recognition Networks. In Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Long Beach, CA, USA, 16–17 June 2019; pp. 1617–1623. [Google Scholar]
  26. Lin, B.; Li, X.; Yu, Z.; Zhao, G. Face Liveness Detection by rPPG Features and Contextual Patch-Based CNN. In Proceedings of the 2019 3rd International Conference on Biometric Engineering and Applications (ICBEA 2019), Stockholm, Sweden, 29–31 May 2019; pp. 61–68. [Google Scholar]
  27. Chen, S.; Liu, Y.; Gao, X.; Han, Z. MobileFaceNets: Efficient CNNs for Accurate Real-Time Face Verification on Mobile Devices. In Biometric Recognition, Proceedings of the CCBR 2018, Urumqi, China, 11–12 August 2018; Zhou, J., Wang, Y., Sun, Z., Jia, Z., Feng, J., Shan, S., Ubul, K., Guo, Z., Eds.; Lecture Notes in Computer Science; Springer: Cham, Switzerland, 2018; Volume 10996. [Google Scholar]
  28. Hou, Q.; Zhou, D.; Feng, J. Coordinate Attention for Efficient Mobile Network Design. In Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Nashville, TN, USA, 19–25 June 2021; pp. 13708–13717. [Google Scholar]
  29. Yu, Z.; Zhao, C.; Wang, Z.; Qin, Y.; Su, Z.; Li, X.; Zhou, F.; Zhao, G. Searching Central Difference Convolutional Networks for Face Anti-Spoofing. In Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, USA, 13–19 June 2020; pp. 5294–5304. [Google Scholar]
  30. Rössler, A.; Cozzolino, D.; Verdoliva, L.; Riess, C.; Thies, J.; Niessner, M. FaceForensics++: Learning to Detect Manipulated Facial Images. In Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision (ICCV), Seoul, Repiblic of Korea, 27 October–2 November 2019; pp. 1–11. [Google Scholar]
  31. Tan, X.; Li, Y.; Liu, J.; Jiang, L. Face Liveness Detection from a Single Image with Sparse Low Rank Bilinear Discriminative Model. In Computer Vision–ECCV 2010, Heraklion, Crete, Greece, 5–11 September 2010; Lecture Notes in Computer Science; Springer: Berlin/Heidelberg, Germany, 2010; Volume 6316. [Google Scholar]
  32. Zhang, Z.; Yan, J.; Liu, S.; Lei, Z.; Yi, D.; Li, S.Z. A Face Antispoofing Database with Diverse Attacks. In Proceedings of the 2012 5th IAPR International Conference on Biometrics (ICB), New Delhi, India, 29 March–1 April 2012; pp. 26–31. [Google Scholar]
  33. Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn, D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.; Heigold, G.; Gelly, S.; et al. An Image is Worth 16 × 16 Words: Transformers for Image Recognition at Scale. arXiv 2020, arXiv:2010.11929. [Google Scholar]
  34. He, K.; Chen, X.; Xie, S.; Li, Y.; Dollár, P.; Girshick, R. Masked Autoencoders Are Scalable Vision Learners. In Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, LA, USA, 18–24 June 2022; pp. 15979–15988. [Google Scholar]
  35. George, A.; Mostaani, Z.; Geissenbuhler, D.; Nikisins, O.; Anjos, A.; Marcel, S. Biometric Face Presentation Attack Detection with Multi-Channel Convolutional Neural Network. IEEE Trans. Inf. Forensics Secur. 2020, 15, 42–55. [Google Scholar] [CrossRef]
  36. Zhang, S.; Liu, A.; Wan, J.; Liang, Y.; Guo, G.; Escalera, S.; Escalante, H.J.; Li, S.Z. CASIA-SURF: A Large-Scale Multi-Modal Benchmark for Face Anti-Spoofing. IEEE Trans. Biom. Behav. Identity Sci. 2020, 2, 182–193. [Google Scholar] [CrossRef]
  37. Liu, A.; Tan, Z.; Wan, J.; Liang, Y.; Lei, Z.; Guo, G.; Li, S.Z. Face Anti-Spoofing via Adversarial Cross-Modality Translation. IEEE Trans. Inf. Forensics Secur. 2021, 16, 2759–2772. [Google Scholar] [CrossRef]
  38. He, K.; Zhang, X.; Ren, S.; Sun, J. Deep Residual Learning for Image Recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 27–30 June 2016; pp. 770–778. [Google Scholar]
  39. Hu, J.; Shen, L.; Sun, G. Squeeze-and-Excitation Networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Salt Lake City, UT, USA, 18–22 June 2018; pp. 7132–7141. [Google Scholar]
  40. Fang, H.; Liu, A.; Wan, J.; Escalera, S.; Zhao, C.; Zhang, X.; Li, S.Z.; Lei, Z. Surveillance Face Anti-Spoofing. IEEE Trans. Inf. Forensics Secur. 2024, 19, 1535–1546. [Google Scholar] [CrossRef]
  41. Li, X.; Komulainen, J.; Zhao, G.; Yuen, P.-C.; Pietikäinen, M. Generalized Face Anti-Spoofing by Detecting Pulse from Face Videos. In Proceedings of the International Conference on Pattern Recognition (ICPR), Cancun, Mexico, 4–8 December 2016; pp. 4244–4249. [Google Scholar]
  42. Wen, D.; Han, H.; Jain, A.K. Face Spoof Detection with Image Distortion Analysis. IEEE Trans. Inf. Forensics Secur. 2015, 10, 746–761. [Google Scholar] [CrossRef]
  43. Erdogmus, N.; Marcel, S. Spoofing Face Recognition with 3D Masks. IEEE Trans. Inf. Forensics Secur. 2014, 9, 1084–1097. [Google Scholar] [CrossRef]
  44. Liu, S.; Yuen, P.C.; Zhang, S.; Zhao, G. 3D Mask Face Anti-spoofing with Remote Photoplethysmography. In Proceedings of the Computer Vision–ECCV 2016, Amsterdam, The Netherlands, 11–14 October 2016; Lecture Notes in Computer Science. Springer: Cham, Switzerland, 2016; Volume 9911. [Google Scholar]
  45. Boulkenafet, Z.; Komulainen, J.; Li, L.; Feng, X.; Hadid, A. OULU-NPU: A Mobile Face Presentation Attack Database with Real-World Variations. In Proceedings of the 2017 12th IEEE International Conference on Automatic Face & Gesture Recognition (FG), Washington, DC, USA, 30 May–3 June 2017; pp. 612–618. [Google Scholar]
  46. Alotaibi, A.; Mahmood, A. Deep Face Liveness Detection Based on Nonlinear Diffusion Using Convolution Neural Network. Signal Image Video Process. 2017, 11, 713–720. [Google Scholar] [CrossRef]
  47. Kim, W.; Suh, S.; Han, J.-J. Face Liveness Detection From a Single Image via Diffusion Speed Model. IEEE Trans. Image Process. 2015, 24, 2456–2465. [Google Scholar] [CrossRef]
  48. Hao, H.; Pei, M.; Zhao, M. Face Liveness Detection Based on Client Identity Using Siamese Network. In Pattern Recognition and Computer Vision, Proceedings of the PRCV 2019, Xi’an, China, 8–11 November 2019; Lecture Notes in Computer Science; Springer: Cham, Switzerland, 2019; Volume 11857. [Google Scholar]
  49. Liu, Y.; Jourabloo, A.; Liu, X. Learning Deep Models for Face Anti-Spoofing: Binary or Auxiliary Supervision. In Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA, 18–22 June 2018; pp. 389–398. [Google Scholar]
  50. Chingovska, I.; Anjos, A.; Marcel, S. On the Effectiveness of Local Binary Patterns in Face Anti-Spoofing. In Proceedings of the 2012 International Conference of Biometrics Special Interest Group (BIOSIG), Darmstadt, Germany, 6–7 September 2012; pp. 1–7. [Google Scholar]
  51. Page, M.J.; McKenzie, J.E.; Bossuyt, P.M.; Boutron, I.; Hoffmann, T.C.; Mulrow, C.D.; Shamseer, L.; Tetzlaff, J.M.; Akl, E.A.; Brennan, S.E.; et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ 2021, 372, n71. [Google Scholar] [CrossRef]
Information 16 00581 g001
Figure 1. The PRISMA flow diagram for face recognition payment system. * No automation tools were used in the screening process, all screening was performed manually by human reviewers.
Figure 1. The PRISMA flow diagram for face recognition payment system. * No automation tools were used in the screening process, all screening was performed manually by human reviewers.
Information 16 00581 g001
Information 16 00581 g002
Figure 2. The number of publications by year and database.
Figure 2. The number of publications by year and database.
Information 16 00581 g002
Table 1. Comparison of face anti-spoofing models for face recognition payments.
Table 1. Comparison of face anti-spoofing models for face recognition payments.
AuthorYearModelTypeDatasetMetrics
Xiao et. al. [17]2024MobileFaceNet + Coordinate Attention (CA)CNNLQFA-DACER: 1.39%
Guo et. al. [18]2024Multiscale Attention Halo Attention Network (MAHA-Net)CNN(a) FF++ HQ
(b) FF++ LQ		(a) ACC: 97.12%
(b) ACC: 91.26%
Yu et. al. [19]2024Vision Transformer (ViT) + Adaptive Multimodal Adapter (AMA) + Modality Asymmetric Masked Autoencoder (M2A2E)ViT(a) WMCA
(b) CASIA-SURF
(c) CeFA		(a) ACER: 6.08%
(b) ACER: 2.12%
(c) ACER: 8.36%
Wang et. al. [20]2023Dynamic Feature Queue (DFQ) + Progressive Training Strategy (PTS)ViTSuHiFiMaskACER: 4.73%
Zou et. al. [21]2023High Frequency Invariant Learning Module (hf-IFLM) + High Dimensional Feature Distribution Learning Module (hd-FDLM)CNN(a) FF++ MQ
(b) FF++ LQ		(a) ACC: 96.56%
(b) ACC: 90.29%
Li et. al. [22]2023Middle Shallow Feature Aggregation (MSFA)CNNCASIA-SURFACER: 0.079%
Wang et. al. [23]2023Siamese NetworkCNN(a) CASIA-FASD
(b) Replay-Attack		(a) HTER: 23.9%
(b) HTER: 38.0%
Lu et. al. [24]2020Heterogeneous Kernel–Convolutional Neural Network (HK-CNN)CNN(a) NUAA
(b) CASIA-FASD		(a) ACC: 99.82%
(b) ACC: 99.17%
Parkin et. al. [25]2019Multi Level Feature Aggregation (MLFA)CNNCASIA-SURFTPR: 99.87%
Lin et. al. [26]2019Multi Scale Long-Term Statistical Spectral (MS-LTSS) + Contextual Patch-based CNN (CP-CNN)CNN(a) MSU-MFSD
(b) CASIA-FASD
(c) OULU-NPU		(a) EER: 0.0%
(b) EER: 1.8%
(c) ACER: 15.1%
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Pratama, M.H.A.; Rizal, A.; Irawati, I.D. Advancing Secure Face Recognition Payment Systems: A Systematic Literature Review. Information 2025, 16, 581. https://doi.org/10.3390/info16070581

AMA Style

Pratama MHA, Rizal A, Irawati ID. Advancing Secure Face Recognition Payment Systems: A Systematic Literature Review. Information. 2025; 16(7):581. https://doi.org/10.3390/info16070581

Chicago/Turabian Style

Pratama, M. Haswin Anugrah, Achmad Rizal, and Indrarini Dyah Irawati. 2025. "Advancing Secure Face Recognition Payment Systems: A Systematic Literature Review" Information 16, no. 7: 581. https://doi.org/10.3390/info16070581

APA Style

Pratama, M. H. A., Rizal, A., & Irawati, I. D. (2025). Advancing Secure Face Recognition Payment Systems: A Systematic Literature Review. Information, 16(7), 581. https://doi.org/10.3390/info16070581

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop