VOIDFace: A Privacy-Preserving Multi-Network Face Recognition With Enhanced Security

2025-08-11

AI Summary

Existing facial recognition systems rely on centralized replication and storage of facial data, leading to privacy breaches, regulatory challenges, and loss of user data control. This paper proposes VOIDFace—the first framework to integrate visual secret sharing (VSS) into facial recognition training—enabling plaintext-free facial data storage and decentralized collaborative model training. Its core innovation is an image-patch-based multi-network architecture that supports secure model updates directly over VSS ciphertexts, while natively supporting the “right to be forgotten”: users can unilaterally revoke authorization, and the system performs secure model erasure without accessing raw data. Experiments on VGGFace2 demonstrate that VOIDFace achieves state-of-the-art recognition accuracy while eliminating redundant data replication entirely, significantly enhancing privacy preservation and reinforcing user data sovereignty.

Technology Category

• Computer Vision: Bias, Fairness & Privacy
• Machine Learning: Privacy
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy

Application Category

• Security and Privacy: Data transparency and provenance
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems

Abstract

Advancement of machine learning techniques, combined with the availability of large-scale datasets, has significantly improved the accuracy and efficiency of facial recognition. Modern facial recognition systems are trained using large face datasets collected from diverse individuals or public repositories. However, for training, these datasets are often replicated and stored in multiple workstations, resulting in data replication, which complicates database management and oversight. Currently, once a user submits their face for dataset preparation, they lose control over how their data is used, raising significant privacy and ethical concerns. This paper introduces VOIDFace, a novel framework for facial recognition systems that addresses two major issues. First, it eliminates the need of data replication and improves data control to securely store training face data by using visual secret sharing. Second, it proposes a patch-based multi-training network that uses this novel training data storage mechanism to develop a robust, privacy-preserving facial recognition system. By integrating these advancements, VOIDFace aims to improve the privacy, security, and efficiency of facial recognition training, while ensuring greater control over sensitive personal face data. VOIDFace also enables users to exercise their Right-To-Be-Forgotten property to control their personal data. Experimental evaluations on the VGGFace2 dataset show that VOIDFace provides Right-To-Be-Forgotten, improved data control, security, and privacy while maintaining competitive facial recognition performance. Code is available at: https://github.com/ajnasmuhammed89/VOIDFace

Problem

• Eliminates data replication in face recognition training
• Enhances privacy and control over personal face data
• Proposes secure patch-based multi-network recognition system

Innovation

• Visual secret sharing for secure face data storage
• Patch-based multi-training network for privacy preservation
• Right-To-Be-Forgotten enabled user data control