Mandadapu Srinadh, Peruri Rupesh, Katakota Hemanth Kumar, Chinagudaba Sai Chetan

DOI: https://doi.org/10.5281/zenodo.23255729

Photographs and videos are shared online within minutes of being recorded, and a single clip often shows people who were never asked whether they agree to appear in it. Manual privacy controls put the burden on the affected person and usually act after the content is already public. This paper presents Safe-Nation Vision, an AI-based identity protection and consent-management system for digital images and videos. Uploaded media is analyzed, faces are located with computer-vision tools, and each face is converted to a facial embedding and compared with the embeddings of registered users. When a registered person is recognized, the system checks the privacy status of the content and sends a consent request. The person answers Allow, Allow Every Time or Don’t Allow, and the media is published or restricted accordingly; unmatched faces are labelled Unknown. The design follows five layers, from user interface to notification and publishing control, and the project followed the Waterfall model. The paper describes the requirements, architecture, workflow and test plan of the system. The paper does not claim measured recognition accuracy or processing time. The contribution is a workflow in which publication depends on the approval of the people who appear in the media.

identity protection, consent management, face detection, face recognition, computer vision, digital privacy, content classification

This paper described Safe-Nation Vision, an AI-based system that ties identity recognition to consent for digital images and videos. Faces are detected, converted to embeddings and matched against registered users. Those users are asked to allow or deny publication, and their answers decide whether the media is published or restricted. This connection between recognition and consent is what distinguishes the system from surveillance and moderation tools that only detect events or harmful content. What the project has established so far is the workflow, the layered architecture, the module and requirement specification, and a test plan. Measured accuracy, latency and test results are not yet available, so the performance targets remain targets. The system does not solve identity misuse as a whole. It protects only registered individuals, depends on recognition quality, and needs production-level security before real use. Future work includes real-time consent validation, deepfake detection, stronger recognition models and privacy-preserving processing.

[1]S. Z. Li and A. K. Jain, “Facial recognition systems for law enforcement applications,” IEEE Computer, vol. 44, no. 1, pp. 40–47, 2022.
[2]R. Cucchiara and C. Grana, “Intelligent video surveillance for public safety,” IEEE Trans. Syst., Man, Cybern., vol. 40, no. 6, pp. 1721–1733, 2010.
[3]Y. Sultani, C. Chen, and M. Shah, “Real-time anomaly detection in surveillance videos,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2018, pp. 6479–6488.
[4]A. Datta and K. Roy, “Deep learning for crime detection in surveillance videos,” Pattern Recognit. Lett., vol. 105, pp. 256–263, 2018.
[5]T. Hassner and Y. Itcher, “Violence detection in surveillance videos using deep neural networks,” Pattern Recognit., vol. 94, pp. 321–331, 2019.
[6]S. Ravanbakhsh and M. Nabi, “Crowd behavior analysis for public safety using deep learning,” IEEE Trans. Intell. Transp. Syst., vol. 19, no. 11, pp. 3564–3575, 2018.
[7]F. Ordóñez and D. Roggen, “Human activity recognition using deep learning for smart surveillance,” Sensors, vol. 16, no. 1, pp. 1–22, 2016.
[8]A. Olmos and J. C. Ferrer, “Gun detection in surveillance cameras using deep learning,” Appl. Sci., vol. 8, no. 9, pp. 1–15, 2018.
[9]R. Girshick and Y. Taigman, “Deep learning-based face detection and recognition in intelligent surveillance systems,” Int. J. Comput. Vis. Pattern Recognit., vol. 18, no. 4, pp. 114–125, 2021.
[10]J. Ren and H. Zhang, “Privacy-preserving smart surveillance using edge AI,” IEEE Internet Things J., vol. 17, no. 9, pp. 8448–8458, 2020.
[11]P. Mach and Z. Becvar, “Edge computing for intelligent surveillance systems,” IEEE Commun. Mag., vol. 55, no. 8, pp. 102–108, 2017.
[12]S. Kumar and P. Patel, “Smart city surveillance using IoT and artificial intelligence,” Int. J. Adv. Comput. Sci. Appl., vol. 11, no. 4, pp. 215–223, 2020.
[13]M. Ahmed and A. Mahmood, “Anomaly detection in smart cities using machine learning,” J. Big Data, vol. 6, no. 1, pp. 1–24, 2019.
[14]J. Salamon and J. P. Bello, “Audio-based anomaly detection for public safety systems,” IEEE Signal Process. Lett., vol. 24, no. 8, pp. 1200–1204, 2017.
[15]K. Chen and L. Sun, “Multimodal sensor fusion for intelligent public safety systems,” Sensors, vol. 20, no. 6, pp. 1742–1754, 2020.
[16]A. Imran and C. Castillo, “Social media mining for real-time crisis detection,” ACM Trans. Inf. Syst., vol. 35, no. 1, pp. 1–28, 2017.
[17]K. Puri and R. Verma, “Drone-based surveillance systems for public safety,” J. Intell. Robot. Syst., vol. 95, no. 3, pp. 567–580, 2019.
[18]A. Gupta and R. Banerjee, “Disaster detection and response using artificial intelligence,” Int. J. Disaster Risk Reduct., vol. 45, Art. no. 101112, 2020.
[19]L. Zhang and H. Wang, “AI-based real-time alert system for public safety,” Future Gener. Comput. Syst., vol. 4, no. 6, pp. 463–472, 2019.
[20]D. Lee and S. Park, “AI-driven national public safety platforms,” Gov. Inf. Q., vol. 37, no. 2, pp. 101–109, 2020.

Mandadapu Srinadh, Peruri Rupesh, Katakota Hemanth Kumar, Chinagudaba Sai Chetan (2026). Safe-Nation Vision: AI-Based Identity Protection and Consent-Driven Digital Content System. International Journal of Computer Science Engineering Techniques, 10(5), 257–265. ISSN: 2455-135X. DOI: https://doi.org/10.5281/zenodo.23255729