Face Recognition

Attend-Go uses on-device and server-side AI to perform rapid facial recognition during live sessions. This allows for seamless, hands-free check-ins without the need for proprietary hardware or physical badges.

1. **Enrollment**: An attendee's face is captured via webcam or uploaded photo. The image is processed to extract a 128-dimensional embedding vector. 2. **Storage**: The embedding is securely stored in a tenant-scoped pgvector database. 3. **Live Match**: During a session, the kiosk continuously scans for faces. When a face is detected, its embedding is compared against enrolled attendees using cosine distance. 4. **Visual Feedback**: The interface immediately highlights matched faces with the attendee's name and records the check-in.