Technical portfolio project
Face Enrollment, Liveness, and Attendance System (FAISS + FaceNet + Amazon APIs)
Enrollment and verification platform that prevents duplicates, validates liveness, and logs auditable attendance at scale.
This entry documents MuFaw research and technical work. It is not presented as a paid client engagement unless the project page explicitly states otherwise.
Project details
What we delivered
Overview
Enrollment and attendance platform with duplicate prevention and liveness verification.
Designed for high-volume environments with compliance-focused audit trails.
Key Features
- Enrollment flow with duplicate detection to prevent repeated identities.
- Face embeddings via FaceNet with FAISS similarity search.
- Liveness verification using Amazon APIs as primary or secondary checks.
- Hybrid verification when local FaceNet confidence is low.
- Auditable attendance logs with manual override and HR or ERP integration.
Architecture & Tech Stack
- Embeddings: FaceNet models deployed on-prem or in containers.
- Search: FAISS for sub-second nearest-neighbor retrieval.
- Liveness and verification: Amazon Rekognition and Liveness APIs.
- Backend: FastAPI services for enrollment and verification.
- Security: encrypted embedding storage, role-based access, audit trails.
Impact / Outcomes
- Reduced proxy and spoofing attempts with multi-factor liveness checks.
- Delivered low-latency attendance marking with reliable auditability.
Explore further