Deep learning-based methods for real face recognition applications
Keywords:
automatic face recognition, deep learningAbstract
Introduction: Automatic face recognition has remained an active research topic in the field of computer vision for several years. It is one of the most widely used biometric techniques, but its effectiveness in practical systems is affected by uncontrolled capture conditions such as low resolution, the use of masks, or the recognition of facial sketches in a criminal investigation.
Objective: To develop a set of algorithms for automatic face recognition based on deep learning techniques, in order to be used in real-world applications in uncontrolled environments.
Methods: The research is supported by the latest advancements in deep learning techniques, which have been adapted and contextualized for practical face recognition applications for security and law-enforcement.
Results: In particular, novel methods are proposed for the tasks of: a) efficient automatic face recognition in videos, b) face recognition in low-resolution videos, c) selection of frames and regions for more effective face recognition, d) classification of soft face attributes, e) automatic face recognition with masked faces, and f) recognition of facial sketches.
Conclusions: The developed algorithms achieve increased effectiveness in the aforementioned tasks compared to similar algorithms in the state-of-the-art and have shown to be effective in different practical environments.
Downloads
References
1. Taskiran M, Kahraman N, Erdem CE. Face recognition: past, present and future (a review). Digital Signal Processing. 2020 Nov. 1;106:102809.
2. Wang M, Deng W. Deep face recognition: a survey. Neurocomputing. 2021 Mar 14;429:215-44.
3. Anwarul S, Dahiya S. A comprehensive review on face recognition methods and factors affecting facial recognition accuracy. In: Proceedings of ICRIC 2019: Recent Innovations in Computing. Cham: Springer; 2020. p. 495-514.
4. Bashbaghi S, Granger E, Sabourin R, Parchami M. Deep learning architectures for face recognition in video surveillance. In: Deep learning in object detection and recognition. Singapore: Springer; 2019. p. 133-54.
5. Liu W, Zhou L, Chen J. Face recognition based on lightweight convolutional neural networks. Information. 2021 Apr 28;12(5):191.
6. Li P, Prieto L, Mery D, Flynn PJ. On low-resolution face recognition in the wild: comparisons and new techniques. IEEE Transactions on Information Forensics and Security. 2019 Jan 3;14(8):2000-12.
7. Alzubi A, Albalas F, Al-Hadhrami T, Younis LB, Bashayreh A. Masked face recognition using deep learning: a review. Electronics. 2021 Oct 31;10(21):2666.
8. Jacquet M, Champod C. Automated face recognition in forensic science: review and perspectives. Forensic Science International. 2020 Feb 1;307:110124.
9. Parkhi O, Vedaldi A, Zisserman A. Deep face recognition. In: Proceedings of the British Machine Vision Conference (BMVC) 2015; 2015 Sep 7-10; Swansea, UK. Swansea: British Machine Vision Association; 2015. p. 41.1-41.12.
10. López-Avila L, Plasencia-Calaña Y, Martínez-Díaz Y, Méndez-Vázquez H. On the use of pre-trained neural networks for different face recognition tasks. In: Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications: 22nd Iberoamerican Congress, CIARP 2017; 2017 Nov 7-10; Valparaíso, Chile. Cham: Springer; 2018. p. 356-64.
11. Martinez-Diaz Y, Luevano LS, Mendez-Vazquez H, Nicolas-Diaz M, Chang L, Gonzalez-Mendoza M. Shufflefacenet: a lightweight face architecture for efficient and highly-accurate face recognition. In: Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops; 2019 Oct 27-Nov 2; Seoul, Korea. Piscataway: IEEE; 2019. p. 0-0.
12. Hernández-Durán M, Plasencia-Calaña Y, Méndez-Vázquez H. Low-resolution face recognition with deep convolutional features in the dissimilarity space. In: Progress in Artificial Intelligence and Pattern Recognition: 6th International Workshop, IWAIPR 2018; 2018 Sep 24-26; Havana, Cuba. Cham: Springer; 2018. p. 95-103.
13. Martinez-Diaz Y, Mendez-Vazquez H, Luevano LS. Effectiveness of blind face restoration to boost face recognition performance at low-resolution images. In: Progress in Artificial Intelligence and Pattern Recognition: 8th International Workshop on Artificial Intelligence and Pattern Recognition (IWAIPR 2023); 2023 Sep 26-28; Varadero, Cuba. Cham: Springer; 2023. p. 455-67.
14. Bahroun S, Abed R, Zagrouba E. KS-FQA: keyframe selection based on face quality assessment for efficient face recognition in video. IET Image Processing. 2021 Jan;15(1):77-90.
15. Méndez-Llanes N, Castillo-Rosado K, Méndez-Vázquez H, Khellat-Kihel S, Tistarelli M. Face recognition on mobile devices based on frames selection. In: Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications: 24th Iberoamerican Congress, CIARP 2019; 2019 Oct 28-31; Havana, Cuba. Cham: Springer; 2019. p. 316-25.
16. Méndez-Llanes N, Castillo-Rosado K, Méndez-Vázquez H, Tistarelli M. Quality-based representation for unconstrained face recognition. In: 2020 25th International Conference on Pattern Recognition (ICPR); 2021 Jan 10-15; Milan, Italy. IEEE; 2021. p. 6494-500.
17. Rivero-Hernandez J, Morales-Gonzalez A, Denis LG, Mendez-Vazquez H. Ordered weighted aggregation networks for video face recognition. Pattern Recognition Letters. 2021 Jun 1;146:237-43.
18. Becerra-Riera F, Morales-González A, Méndez-Vázquez H. A survey on facial soft biometrics for video surveillance and forensic applications. Artificial Intelligence Review. 2019 Aug;52(2):1155-87.
19. Becerra-Riera F, Morales-Gonzalez A, Mendez-Vazquez H, Dugelay JL. Attribute-based quality assessment for demographic estimation in face videos. In: 2020 25th International Conference on Pattern Recognition (ICPR); 2021 Jan 10-15; Milan, Italy. IEEE; 2021. p. 5875-82.
20. Becerra-Riera F, Morales-González A, Méndez-Vázquez H, Dugelay JL. Demographic attribute estimation in face videos combining local information and quality assessment. Machine Vision and Applications. 2022 Mar;33(2):26.
21. Becerra-Riera F, Morales-González A, Méndez-Vázquez H. Exploring local deep representations for facial gender classification in videos. In: Progress in Artificial Intelligence and Pattern Recognition: 6th International Workshop, IWAIPR 2018; 2018 Sep 24-26; Havana, Cuba. Cham: Springer; 2018. p. 104-12.
22. Deng J, Guo J, An X, Zhu Z, Zafeiriou S. Masked face recognition challenge: the insightface track report. In: Proceedings of the IEEE/CVF International Conference on Computer Vision; 2021 Oct 11-17; Virtual. IEEE; 2021. p. 1437-44.
23. Martínez-Díaz Y, Méndez-Vázquez H, Luevano LS, Nicolás-Díaz M, Chang L, González-Mendoza M. Towards accurate and lightweight masked face recognition: an experimental evaluation. IEEE Access. 2021 Dec 13;10:7341-53.
24. Méndez-Vázquez H, Becerra-Riera F, Morales-González A, López-Avila L, Tistarelli M. Local deep features for composite face sketch recognition. In: 2019 7th International Workshop on Biometrics and Forensics (IWBF); 2019 May 2-3; Cancun, Mexico. IEEE; 2019. p. 1-6.
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 Heydi Mendez Vazquez
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
The journal Anales de la Academia de Ciencias de Cuba protects copyright, and operates with a Creative Commons License 4.0 (Creative Commons Attribution-NonCommercial License 4.0). By publishing in it, authors allow themselves to copy, reproduce, distribute, publicly communicate their work and generate derivative works, as long as the original author is cited and acknowledged. They do not allow, however, the use of the original work for commercial or lucrative purposes.
The authors authorize the publication of their writings, retaining the authorship rights, and assigning and transferring to the magazine all the rights protected by the intellectual property laws that govern in Cuba, which imply editing to disseminate the work.
Authors may establish additional agreements for the non-exclusive distribution of the version of the work published in the journal (for example, placing it in an institutional repository or publishing it in a book), with recognition of having been first published in this journal.
To learn more, see https://creativecommons.org
