Face Recognition Attendance System

The challenge

The client used card-based punching across several sites. Cards were shared or lost, "buddy punching" was common, and HR spent days every month reconciling attendance before running payroll. Field staff had no reliable way to mark attendance at all.

Our solution

We delivered an AI-powered face recognition attendance system:

• Face recognition: employees mark attendance in about a second using a wall-mounted tablet or their own phone.
• Liveness detection: blocks photos and videos, so attendance cannot be faked.
• Geo-fencing: mobile check-ins are only accepted within the approved site location, which suits field and remote staff.
• Works offline: kiosks keep working without internet and sync later.
• Shifts and rules: shift rosters, late marks, overtime, leave and holiday rules.
• Reports and payroll: daily muster, monthly reports and export or API integration with the payroll system.
• Privacy first: face data is stored as encrypted templates rather than photos, with consent and role-based access.

Technology: Python, deep-learning face recognition models, liveness detection, Flutter, Node.js, PostgreSQL.

Results

• Buddy punching eliminated, because each check-in is matched to a live face
• Contactless and fast at shift change
• Attendance data flows straight into payroll, cutting month-end reconciliation
• One dashboard for every site and for field staff