Manual roll calls and biometric fingerprint scanners caused extensive bottlenecks and hygiene concerns across multi-building academic campuses. Elegensoft engineered a touchless, high-throughput facial recognition pipeline integrated directly into the university’s Enterprise LMS.
The Enterprise Challenge
IIIT Kalyani faced severe logistical hurdles during morning lecture transitions. Over 2,500 students arriving across multiple lecture halls created queuing delays that routinely cost 10–15 minutes per lecture. Existing biometric fingerprint hardware suffered from hardware degradation, latency spikes, and hygiene friction.
Our Architectural Solution
We designed an edge-accelerated computer vision architecture deployed across ceiling-mounted wide-angle IP cameras. The system utilizes lightweight YOLOv8 models for instantaneous multi-face bounding box detection, coupled with FaceNet embeddings running on NVIDIA Jetson edge nodes.
- Real-Time Edge Inference: Sub-80ms facial landmark extraction with zero cloud network dependency.
- Liveness Anti-Spoofing: Dual-spectrum infrared and micro-motion heuristics preventing photo/video playback spoofing.
- Automated LMS Synchronization: Webhook triggers immediately dispatching authenticated attendance rolls into the Moodle LMS.
“The transition to Elegensoft’s AI attendance platform eradicated lecture delays entirely. It handles our peak morning rushes effortlessly with flawless accuracy.”
Dean of Academic Affairs, IIIT Kalyani
Measurable Business & Operational Impact
Within the first academic quarter of deployment, administrative overhead decreased by 80%, recovering over 350 faculty lecture hours. Recognition accuracy under varied lighting conditions reached 99.5%, setting a new standard for campus smart infrastructure.