DISCUSSION ON LARGE-SCALE FACIAL RECOGNITION USING DEEP NEURAL NETWORKS IN THE SADE SYSTEM – PMPR
Abstract
This article presents a methodological and experimental study on the use of deep learning-based facial recognition techniques in a hypothetical scenario inspired by institutional public security systems, focusing on the possible integration of a biometric module into SADE (Emergency Response and Dispatch System) of the Military Police of Paraná as a decision-support tool. In contrast to approaches aimed at immediate deployment, the study critically examines the technical feasibility, operational limitations, and ethical, legal, and organizational implications associated with the use of this technology in sensitive contexts. The results show that the developed system was able to organize a facial embedding space with consistent discriminative power, indicating that performance in identification tasks depends directly on the quantity and diversity of images available for each identity in the gallery. In verification evaluations, stricter security settings were found to reduce false positives, but also to increase the rejection rates of genuine individuals. This finding reinforces that aggregated accuracy metrics, when considered in isolation, are insufficient to validate the use of this technology in police applications. From a methodological perspective, the study contributes by describing a reproducible pipeline for data organization, model training, embedding generation, and experimental evaluation. Based on the findings, any potential integration into SADE should be conceived as a probabilistic screening service, always subject to human oversight, process auditability, institutional governance, and the protection of fundamental rights.
Author Biography
Technologist in Artificial Intelligence, Cruzeiro do Sul University, São Paulo-SP, Brazil. Enlisted Personnel of the Military Police of the State of Paraná, Cornélio Procópio, Paraná, Brazil.
References
ALMEIDA, Lucas A.; SILVA, Renata M.; COSTA, Henrique P. Reconhecimento facial, segurança pública e direitos fundamentais: desafios éticos e jurÃdicos. Revista de Ciências Criminais Contemporâneas, São Paulo, v. 8, n. 2, p. 45–63, 2021.
BRANDNER, Tobias. Facial recognition technologies: opportunities, risks and ethical implications. AI & Society, London, v. 38, n. 1, p. 125–138, 2023 DOI: https://doi.org/10.1007/s00146-022-01465-9
BRASIL. Lei nº 13.675, de 11 de junho de 2018. Institui o Sistema Único de Segurança Pública (Susp) e cria a PolÃtica Nacional de Segurança Pública e Defesa Social (PNSPDS). BrasÃlia, DF: Presidência da República, 2018. DisponÃvel em: https://www.planalto.gov.br/ccivil_03/_ato2015-2018/2018/lei/L13675.htm Acesso em: 1 mar. 2026.
BRASIL. Lei nº 13.709, de 14 de agosto de 2018. Dispõe sobre a proteção de dados pessoais e altera a Lei nº 12.965, de 23 de abril de 2014 (Marco Civil da Internet). BrasÃlia, DF: Presidência da República, 2018. DisponÃvel em: https://www.planalto.gov.br/ccivil_03/_ato2015-2018/2018/lei/l13709.htm Acesso em: 1 mar. 2026.
BUOLAMWINI, Joy; GEBRU, Timnit. Gender Shades: intersectional accuracy disparities in commercial gender classification. In: CONFERENCE ON FAIRNESS, ACCOUNTABILITY AND TRANSPARENCY, 2018, New York. Proceedings [...] New York: PMLR, 2018. p. 77–91.
CAO, Qiong; SHEN, Li; XIE, Weidi; PARKHI, Omkar M.; ZISSERMAN, Andrew. VGGFace2: a dataset for recognising faces across pose and age. In: IEEE INTERNATIONAL CONFERENCE ON AUTOMATIC FACE & GESTURE RECOGNITION, 2018, Xi’an, China. Proceedings [...] [S. l.]: IEEE, 2018. p. 67–74.
CENTER ON PRIVACY & TECHNOLOGY. Garbage in, garbage out: face recognition on flawed data. Washington, DC: Georgetown University Law Center, 2019. DisponÃvel em: https://www.law.georgetown.edu/privacy-technology-center/publications/garbage-in-garbage-out-face-recognition-on-flawed-data/ Acesso em: 29 jan. 2026.
CRAWFORD, Kate. Atlas of AI: power, politics, and the planetary costs of artificial intelligence. New Haven: Yale University Press, 2021.
DENG, Jiankang; GUO, Jia; XUE, Niannan; ZAFEIRIOU, Stefanos. ArcFace: additive angular margin loss for deep face recognition. In: IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, 2019, Long Beach, CA. Proceedings [...] [S. l.]: IEEE, 2019. p. 4690–4699.
DONEDA, Danilo et al. Proteção de dados pessoais: a função e os limites do consentimento. Rio de Janeiro: Forense, 2019.
EUROPEAN UNION AGENCY FOR FUNDAMENTAL RIGHTS. Facial recognition technology: fundamental rights considerations in the context of law enforcement. Luxembourg: Publications Office of the European Union, 2019. DisponÃvel em: https://fra.europa.eu/en/publication/2019/facial-recognition-technology-fundamental-rights-considerations-context-law Acesso em: 17 fev. 2026.
GARVIE, Clare; BEDOYA, Alvaro; FRANKLE, Jonathan. The perpetual line-up: unregulated police face recognition in America. Washington, DC: Georgetown Law Center on Privacy & Technology, 2016.
GROTHER, Patrick; NGAN, Mei; HANAOKA, Kayee. Face recognition vendor test (FRVT): demographic effects. Gaithersburg, MD: National Institute of Standards and Technology, 2019 DOI: https://doi.org/10.6028/NIST.IR.8280
INTRONA, Lucas D.; WOOD, David. Picturing algorithmic surveillance: the politics of facial recognition systems. Surveillance & Society, London, v. 2, n. 2/3, p. 177–198, 2004 DOI: https://doi.org/10.24908/ss.v2i2/3.3373
JAIN, Anil K.; ROSS, Arun; PRABHAKAR, Salil. An introduction to biometric recognition. IEEE Transactions on Circuits and Systems for Video Technology, New York, v. 14, n. 1, p. 4–20, 2004 DOI: https://doi.org/10.1109/TCSVT.2003.818349
KISHORE, Saritha et al. Evaluation of Deep Learning Methods in Face Recognition: datasets, metrics, and results. International Journal on Science and Technology (IJSAT), v. 16, n. 4, out./dez. 2025. DisponÃvel em: https://www.ijsat.org/papers/2025/4/9116.pdf Acesso em: 17 fev. 2026 DOI: https://doi.org/10.71097/IJSAT.v16.i4.9087
MIN, J.; FLYNN, P.; BOWYER, K. W. Using multiple gallery and probe images per person to improve performance of face recognition. In: IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS, 2003, Washington, DC. Proceedings [...] Washington, DC: IEEE, 2003. p. 25 DOI: https://doi.org/10.1109/CVPRW.2003.10033
MIRANDA, João Vitor Arnas de; LIMA, Luis Henrique de. O uso de tecnologias como forma de melhorar a eficiência na segurança pública: um quadro comparativo na PMPR pré e pós implementação do sistema de atendimento e despacho de emergências (SADE). Revista CientÃfica de Alto Impacto, v. 27, n. 127, out. 2023. DisponÃvel em: https://revistaft.com.br/o-uso-de-tecnologias-como-forma-de-melhorar-a-eficiencia-na-seguranca-publica-um-quadro-comparativo-na-pmpr-pre-e-pos-implementacao-do-sistema-de-atendimento-e-despacho-de-emergencias-sade/ Acesso em: 17 fev. 2026 DOI: https://doi.org/10.5281/zenodo.8433291
PARKHI, Omkar M.; VEDALDI, Andrea; ZISSERMAN, Andrew. Deep face recognition. In: BRITISH MACHINE VISION CONFERENCE, 2015, Swansea, UK. Proceedings [...] Durham: BMVA Press, 2015. p. 41.1–41.12 DOI: https://doi.org/10.5244/C.29.41
RAPOSO, Vera Lúcia. When facial recognition does not ‘recognise’: erroneous identifications and resulting liabilities. AI & Society, v. 39, p. 1857–1869, 2024 DOI: https://doi.org/10.1007/s00146-023-01634-z
SANTOS, Thiago Federovicz Mendes dos. A implantação do sistema de atendimento e despacho de emergências via mobile no 10º batalhão da PolÃcia Militar do Paraná. Brazilian Journal of Development, Curitiba, v. 9, n. 3, p. 11214–11241, 2023 DOI: https://doi.org/10.34117/bjdv9n3-153
SCHEIRER, Walter J.; JAIN, Lalit P.; BOULT, Terrance E. Probability models for open set recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, New York, v. 36, n. 11, p. 2317–2324, 2014 DOI: https://doi.org/10.1109/TPAMI.2014.2321392
SCHROFF, Florian; KALENICHENKO, Dmitry; PHILBIN, James. FaceNet: a unified embedding for face recognition and clustering. In: IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, 2015, Boston, MA. Proceedings [...] [S. l.]: IEEE, 2015. p. 815–823 DOI: https://doi.org/10.1109/CVPR.2015.7298682
SU, Yiyang; KIM, Minchul; LIU, Feng; JAIN, Anil; LIU, Xiaoming. Open-Set Biometrics: Beyond Good Closed-Set Models. arXiv, [cs.CV], 2024. DisponÃvel em: https://arxiv.org/abs/2407.16133 Acesso em: 17 fev. 2026.
TAIGMAN, Yaniv; YANG, Ming; RANZATO, Marc’Aurelio; WOLF, Lior. DeepFace: closing the gap to human-level performance in face verification. In: IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, 2014, Columbus, OH. Proceedings [...] [S. l.]: IEEE, 2014. p. 1701–1708.
THORAT, S. B.; NAYAK, S. K.; DANDALE, Jyoti P. Facial recognition technology: an analysis with scope in India. International Journal of Computer Science and Information Security, v. 8, n. 1, 2010. DisponÃvel em: https://arxiv.org/abs/1005.4263 Acesso em: 17 fev. 2026.
WANG, Xukang; WU, Ying Cheng; ZHOU, Mengjie; FU, Hongpeng. Beyond surveillance: privacy, ethics, and regulations in face recognition technology. Frontiers in Big Data, Lausanne, v. 7, art. 1337465, 2024. DisponÃvel em: https://www.frontiersin.org/journals/big-data/articles/10.3389/fdata.2024.1337465/full Acesso em: 29 jan. 2026 DOI: https://doi.org/10.3389/fdata.2024.1337465
WASHINGTON POST. Arrested by AI: police ignore standards after facial recognition matches. Washington, DC, 2025. DisponÃvel em: https://www.washingtonpost.com/business/interactive/2025/police-artificial-intelligence-facial-recognition/ Acesso em: 29 jan. 2026.
License
Copyright (c) 2026 RECIMA21 - Revista CientÃfica Multidisciplinar - ISSN 2675-6218
This work is licensed under a Creative Commons Attribution 4.0 International License.
Os direitos autorais dos artigos/resenhas/TCCs publicados pertecem à revista RECIMA21, e seguem o padrão Creative Commons (CC BY 4.0), permitindo a cópia ou reprodução, desde que cite a fonte e respeite os direitos dos autores e contenham menção aos mesmos nos créditos. Toda e qualquer obra publicada na revista, seu conteúdo é de responsabilidade dos autores, cabendo a RECIMA21 apenas ser o veÃculo de divulgação, seguindo os padrões nacionais e internacionais de publicação.Â