A Context-Aware and Explainable Model for Privacy Protection in Social Networks

Raghad S. Mohamed

doi:10.61704/pr.719

المؤلفون

Raghad S. Mohamed Department of Software, College of Computer Science and Mathematics, University of Mosul, Ninevah 41001, Iraq https://orcid.org/0009-0007-1434-2342

DOI:

https://doi.org/10.61704/pr.719

الكلمات المفتاحية:

Social Network Privacy، Explainable Artificial Intelligence، Context-Aware Privacy، Image Privacy Protection، Preserving AI

الملخص

With the rapid growth of social networking platforms, users share their content online that can potentially compromise their privacy. Images shared on social networks can expose information such as their faces, identity documents, locations, health conditions, screens, license plates, children, private environments, and more. While various approaches to the field of social network image privacy-preserving, such as access control, encryption, face blurring, adversarial perturbation, and deep learning-based privacy prediction have been proposed, each approach suffers from three major limitations: insufficient context awareness, weak explainability, and poor balance between privacy protection and content utility. To address these limitations, in this paper, we propose CA-XPrivNet: a context-aware and explainable framework for the privacy protection of content on social networks. CA-XPrivNet includes spatial-domain and frequency-domain methods to sanitize privacy of the shared images while preserving the content that is not sensitive to the social network user. Furthermore, a multimodal risk estimation module determines the risk of the image to reveal sensitive information, and the explainability of the framework is achieved through attention maps, Grad-CAM methods, SHAP methods, and privacy-preserving explanation filtering. To evaluate the proposed framework, experiments were conducted using three publicly available social-network image privacy datasets: PrivacyAlert, DIPA, and DIPA2. These datasets provide a variety of privacy-sensitive images that can be used to evaluate the performance of the framework in terms of its ability to classify privacy risks, reduce privacy leakage, preserve image quality, provide high-quality explanations, and provide fast processing times while providing a balance between privacy and content usability. The results of the experiments indicate that the framework can provide adaptive and explainable privacy protection while preserving the usability of non-sensitive content.

المراجع

Ahmed, K. S., & Baroud, A. A. S. (2022). Protecting Privacy with Image Processing: How Blurring and Masking Help Keep Us Safe Online. Journal for Research in Applied Sciences and Biotechnology, 1(1), 132–138. https://doi.org/10.55544/jrasb.1.1.18

Alzubaidy, H. K., Al-Shammary, D., & Abed, M. H. (2022). A survey on patients privacy protection with stganography and visual encryption [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2201.09388

B, D., Rupa, C., D, H., & Y, V. (2022). Privacy protection of digital information using frequency domain watermarking technique. In 2021 4th International Conference on Recent Trends in Computer Science and Technology (ICRTCST) (pp. 202–206). IEEE. https://doi.org/10.1109/ICRTCST54752.2022.9781929

Baranouskaya, D., & Cavallaro, A. (2023). Human-interpretable and deep features for image privacy classification. In 2023 IEEE International Conference on Image Processing (ICIP) (pp. 3489–3492). IEEE. https://doi.org/10.1109/ICIP49359.2023.10222833

Beigi, G., & Liu, H. (2020). A survey on privacy in social media: Identification, mitigation, and applications. ACM/IMS Transactions on Data Science, 1(1), Article 7. https://doi.org/10.1145/3343038

Cavallaro, A., Malekzadeh, M., & Shamsabadi, A. S. (2020). Deep learning for privacy in multimedia. In Proceedings of the 28th ACM International Conference on Multimedia (pp. 4777–4778). Association for Computing Machinery. https://doi.org/10.1145/3394171.3418551

Chelani, K., Benbihi, A., Kahl, F., Sattler, T., & Kukelova, Z. (2025). Obfuscation based privacy preserving representations are recoverable using neighborhood information. In 2025 International Conference on 3D Vision (3DV) (pp. 189–199). IEEE. https://doi.org/10.1109/3DV66043.2025.00023

Dwivedi, A., Yadav, M., & Kumar, A. (2021). FFT-based zero-bit watermarking for facial recognition and its security. In M. Dave, R. Garg, M. Dua, & J. Hussien (Eds.), Proceedings of the International Conference on Paradigms of Computing, Communication and Data Sciences: PCCDS 2020 (pp. 403–417). Springer. https://doi.org/10.1007/978-981-15-7533-4_31

El Ouadrhiri, A., & Abdelhadi, A. (2022). A new dimensionality reduction method based on Hensel’s compression for privacy protection in federated learning [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2205.02089

El Ouadrhiri, A., Abdelhadi, A., & Phung, P. H. (2023). Hensel’s compression-based dimensionality reduction approach for privacy protection in federated learning. In 2023 International Conference on Computing, Networking and Communications (ICNC) (pp. 298–303). IEEE. https://doi.org/10.1109/ICNC57223.2023.10074197

Fernandez-de-Retana, M., Zulaika, U., Sánchez-Corcuera, R., & Almeida, A. (2025). Differential privacy: Gradient leakage attacks in federated learning environments [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2510.23931

Gupta, A., Jaiswal, A., Wu, Y., Yadav, V., & Natarajan, P. (2021). Adversarial mask generation for preserving visual privacy. In 2021 16th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2021) (pp. 1–5). IEEE. https://doi.org/10.1109/FG52635.2021.9666933

Hassanpour, A., & Yang, B. (2022). PriMe: A novel privacy measuring framework for online social networks. In 2022 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM) (pp. 518–525). IEEE. https://doi.org/10.1109/ASONAM55673.2022.10068701

Huang, Y., Hou, J., Zhao, C., Zhang, J., Xiao, T., & Li, X.-Y. (2025). SCOPE: Bridging explicit and implicit privacy leakage for quantitative image privacy evaluation. IEEE Transactions on Information Forensics and Security, 20, 11560–11575. https://doi.org/10.1109/TIFS.2025.3620117

Ibrahim, A., Montgomery, C., Blackwood, J., & Vance, S. (2026). Privacy-aware visual computing for image analysis in social networks. International Journal of Information and Communication Technology Trends, 6(1), 121–134. https://doi.org/10.66525/ijictt80

Kurakin, A., Song, S., Chien, S., Geambasu, R., Terzis, A., & Thakurta, A. (2022). Toward training at ImageNet scale with differential privacy [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2201.12328

Lange, L., Heykeroth, M.-M., & Rahm, E. (2024). Assessing the impact of image dataset features on privacy-preserving machine learning [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2409.01329

Liu, C., Zhu, T., Zhang, J., & Zhou, W. (2020). Privacy intelligence: A survey on image privacy in online social networks [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2008.12199

Liu, Z., Yang, M. L., & Yan, W. Q. (2021). A framework for image encryption on frequency domain. In Information Resources Management Association (Ed.), Research anthology on artificial intelligence applications in security (pp. 328–338). IGI Global. https://doi.org/10.4018/978-1-7998-7705-9.ch016

Mahloujifar, S., Kokhlikyan, N., Guo, C., & Chaudhuri, K. (2025). Privacy blur: Quantifying privacy and utility for image data release [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2512.16086

Mi, Y., Huang, Y., Ji, J., Zhao, M., Wu, J., Xu, X., Ding, S., & Zhou, S. (2023). Privacy-preserving face recognition using random frequency components. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) (pp. 19616–19627). IEEE. https://doi.org/10.1109/ICCV51070.2023.01802

Montenegro, H., Silva, W., Gaudio, A., Fredrikson, M., Smailagic, A., & Cardoso, J. S. (2022). Privacy-preserving case-based explanations: Enabling visual interpretability by protecting privacy. IEEE Access, 10, 28333–28347. https://doi.org/10.1109/ACCESS.2022.3157589

Sabatino, P., Kumar, C., & Singh, P. (2023). Privacy-preserving image processing with neural networks. World Journal of Advanced Research and Reviews, 18(3), 1686–1693. https://doi.org/10.30574/wjarr.2023.18.3.0690

Salnitri, M., Bonutto, D., Christantoni, I., Kosmidis, D., & Micucci, F. (2020). Visual privacy management: Design and applications of a privacy-enabling platform – Evaluation results (Version 1) [Data set]. Mendeley Data. https://doi.org/10.17632/vs8m5xvbm9.1

Sharma, N., Kumar, R., & Swarnkar, M. (2023). Social networking: Challenges, security issues, and emerging trends. In R. S. Bhadoria, N. Saxena, & B. Nagpal (Eds.), Blockchain technology for secure social media computing (pp. 59–79). The Institution of Engineering and Technology. https://doi.org/10.1049/PBSE019E_ch4

Shibata, H., Hanaoka, S., Koshino, S., Miki, S., Sonoda, Y., & Abe, O. (2024). Identity diffuser: Preserving abnormal region of interests while diffusing identity. Applied Sciences, 14(18), Article 8489. https://doi.org/10.3390/app14188489

Tang, S., Wang, Y., Meng, T., Qin, J., & Lin, L. (2024). Privacy protection for image sensors using convolutional neural network and generative adversarial nets. In 2024 10th International Conference on Computer and Communications (ICCC) (pp. 1192–1198). IEEE. https://doi.org/10.1109/ICCC62609.2024.10942269

Tong, C., Zhang, M., Lang, C., & Zheng, Z. (2021). An image privacy protection algorithm based on adversarial perturbation generative networks. ACM Transactions on Multimedia Computing, Communications, and Applications, 17(2), Article 43. https://doi.org/10.1145/3381088

Tonge, A., & Caragea, C. (2020). Image privacy prediction using deep neural networks. ACM Transactions on the Web, 14(2), Article 7. https://doi.org/10.1145/3386082

Tran, B., Wang, H., Zhang, X., & Jiang, P. (2025). Exploring safer image sharing: A vision-language approach to privacy risk detection and protection. In 2025 IEEE 22nd International Conference on Mobile Ad-Hoc and Smart Systems (MASS) (pp. 537–542). IEEE. https://doi.org/10.1109/MASS66014.2025.00087

Vasantrao, B. S., Sharma, P., & Kadam, S. (2024). AI-driven hybrid privacy preservation model for enhanced security in online social networks. Revista Electronica de Veterinaria, 25(1), 2399–2408. https://doi.org/10.69980/redvet.v25i1.1198

Wang, C., Yang, S., Sun, X., Dai, J., & Zhao, D. (2024). Enabling practical and privacy-preserving image processing [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2409.03568

Wang, Y., Liu, J., Luo, M., Yang, L., & Wang, L. (2022). Privacy-preserving face recognition in the frequency domain. Proceedings of the AAAI Conference on Artificial Intelligence, 36(3), 2558–2566. https://doi.org/10.1609/aaai.v36i3.20157

Wang, Z., Guo, H., Zhang, Z., Song, M., Zheng, S., Wang, Q., & Niu, B. (2020). Towards compression-resistant privacy-preserving photo sharing on social networks. In Proceedings of the Twenty-First International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Networks and Mobile Computing (pp. 81–90). Association for Computing Machinery. https://doi.org/10.1145/3397166.3409141

Zhang, J., Chen, Y., & Li, H. (2022). Privacy leakage of adversarial training models in federated learning systems. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) (pp. 107–113). IEEE. https://doi.org/10.1109/CVPRW56347.2022.00021

Zhao, C., Mangat, J., Koujalgi, S., Squicciarini, A., & Caragea, C. (2022). PrivacyAlert: A dataset for image privacy prediction. Proceedings of the International AAAI Conference on Web and Social Media, 16(1), 1352–1361. https://doi.org/10.1609/icwsm.v16i1.19387