FaceGroup: Continual Face Authentication

​

Paradigm for FaceGroup system: enrollment, group authentication & individual authentication.

Invention Summary:

‘Authentication’ rather than just ‘trust’ is now a changing paradigm for many users and service providers and thus the concept of Zero Trust (ZT) emerged. This for sure, provides enhanced security level for network architectures. Although there are many authentication methods for physical access control, the methods suitable for continual authentication should be able to be conducted remotely, preferably cloud-based.

Rutgers researchers have developed  cloud‐based facial authentication access control solutions to bring ZT into network architectures. Many of the state‐of‐art works on encrypted cloud‐based authentication use Fully Homomorphic Encryption (FHE) to protect users’ private data. Rutgers novel approaches incorporate Partially Homomorphic Encryption (PHE) into cloud‐based two-stage architecture facial authentication (FaceGroup) by changing the distance metric from Euclidean distance to Manhattan distance.

Market Applications:

• Facial verification cuts cost of calculating facial recognition by up to 55%.
• The two-stage architecture for group facial recognition further reduces the total computation cost of authentications.
• FaceGroup lightweight design is scalable and can be deployed on resource-constrained devices.

Advantages:

• PHE is more than capable of processing real-time comparisons of facial feature matrices.
• Manhattan distance demonstrated potential in machine learning for its model training, computation cost, and flexibility.

Intellectual Property & Development Status: Patent pending. Available for licensing and/or research collaboration. For any business development and other collaborative partnerships contact:  marketingbd@research.rutgers.edu