Computer vision / Workflow
FaceAttend
An offline-first face-recognition attendance system with browser-based capture and Excel records intended to make setup and review straightforward.
Context
The project aimed to make attendance capture usable without a database installation. The repository describes webcam recognition, user registration, daily attendance records, and an Excel workbook for reviewing data.
What I worked on
- Developed a workflow for registration and real-time attendance capture.
- Used Excel sheets for user and attendance records rather than requiring a separate database setup.
- Worked on face-detection thresholds after investigating false-positive matches.
- Designed the project to run locally; the project README describes recognition as offline and says photos are not stored.
Implementation notes
The repository separates a Node.js backend from a browser frontend. The front end handles camera-based recognition with face-api.js, while the backend supports workbook-based records. A start script launches the local application.
Scope & evidence
The portfolio summary is based on the resume and the public repository README. No accuracy, adoption, or comparative performance figures are claimed here.