AI / ResearchAIUB — Undergraduate Thesis2022
Attendance Monitoring by Face Recognition
Undergraduate thesis on automated attendance using deep learning and computer vision.
Overview
My undergraduate thesis at AIUB: an attendance monitoring system that identifies students from a classroom camera feed using deep learning based face recognition, removing the manual roll call entirely.
The work covered the full pipeline — face detection and alignment, embedding generation, identity matching against an enrolled set, and the handling of the practical failure cases that make classroom conditions harder than benchmark datasets: occlusion, varied lighting and off-axis faces.
Key work
- 01Deep learning face detection, alignment and embedding pipeline
- 02Identity matching against an enrolled student set
- 03Robustness handling for occlusion, lighting and pose variation
- 04Full academic write-up and defence