Security of Sensitive Data in Face Recognition

Face recognition has emerged as a prominent biometric authentication technique and is increasingly adopted in security-critical domains such as mobile banking, digital identity verification, and online financial transactions. Despite its widespread adoption, ensuring the security and privacy of facial data remains a major challenge due to risks associated with unauthorized access, data breaches, and identity theft. To address these concerns, this project proposes a novel and secure framework that integrates DeepFace, a deep learning–based face recognition library, with blockchain technology to safeguard facial embeddings during storage and transmission. In the proposed system, facial features are extracted from input images using DeepFace and securely stored on a blockchain network in the form of cryptographically hashed and immutable records. The decentralized nature of blockchain ensures data integrity, transparency, and resistance to tampering, while smart contracts regulate access control and authentication processes. A dynamic session management mechanism is incorporated to generate unique transaction identifiers for each authentication request, thereby reducing vulnerability to replay attacks and unauthorized manipulation. The framework is evaluated on a benchmark facial image dataset to assess both recognition performance and security effectiveness. Experimental results demonstrate that the blockchain-based security mechanism introduces minimal computational overhead and does not significantly affect recognition accuracy, achieving an average accuracy rate of 90.1%. Furthermore, the proposed system exhibits constant time complexity, O(1), for verification processes, enabling efficient real-time deployment in resource-constrained environments. By combining robust biometric recognition with decentralized blockchain security, the proposed approach effectively addresses critical privacy, integrity, and trust challenges in modern face recognition systems. Consequently, this framework provides a reliable, transparent, and scalable solution for secure authentication in mobile devices, banking platforms, and other sensitive applications where data confidentiality and user trust are paramount.

Keywords: Face Recognition, Deep Learning, DeepFace, Biometric Authentication, Blockchain Technology, Decentralized Security, Facial Embeddings, Data Integrity, Privacy Preservation, Secure Storage, Smart Contracts, Mobile Authentication, Real-Time Systems, Distributed Ledger, Access Control, Information Security.