| Availability |
Odoo Online
Odoo.sh
On Premise
|
| Odoo Apps Dependencies |
•
Attendances (hr_attendance)
• Employees (hr) • Discuss (mail) |
| Lines of code | 1307 |
| Technical Name |
face_attendance_3d |
| License | LGPL-3 |
3D Face Attendance
AI face recognition & active liveness for Odoo HR Attendance
Let employees clock in and out simply by showing their face. Powered by an external AI service with anti-spoofing (active liveness) checks, so a photo or a screen will not pass.
Key Features
Face Enrollment
Each employee captures several face samples in different poses directly from the browser. Samples are converted to embeddings by the AI service and stored on the employee record.
Verification on Check-in/out
The attendance kiosk verifies the live camera frame against the enrolled employee before recording the attendance.
Active Liveness
When passive checks are not confident enough, the user is asked to follow a randomly generated head-turn sequence, defeating photo and screen spoofing attacks.
Server-side Enforcement
Attendance is only accepted for the logged-in user after a valid, time-limited face verification token, preventing API bypass.
GPS Distance Constraint
Optionally ensure the employee is within an allowed distance from the company location when clocking in or out.
Self-service Only
Employees can only enroll or update their OWN face — even HR managers cannot enroll on behalf of others — guaranteeing the authenticity of attendance.
Configuration
Set the AI service connection in Settings > Technical > System Parameters:
face_attendance_3d.face_ai_url— base URL of the AI service (e.g. http://localhost:8000)face_attendance_3d.face_ai_token— service token sent via theX-Service-Tokenheader (optional)
Note: This module requires a compatible external
face-recognition service exposing /embed,
/compare and /liveness/active endpoints.
Get the AI Service (API & Token)
The face-recognition AI service is provided separately. To obtain the API endpoint URL and the service token, please get in touch with the author on LinkedIn:
Nguyễn Phúc Mạnh — LinkedIn
Please log in to comment on this module