How to Set Up Face Recognition Check-In: Parameters, Liveness Detection, and Face Data Enrollment

In addition to NFC card check-in, TimeClock also supports face recognition check-in.

If employees frequently forget their cards, or if managing physical cards is more trouble than it's worth, face recognition is a clean alternative for fixed-location attendance.

Employees simply step in front of the device's camera, and once the system identifies them, their attendance is recorded. This works well in offices, retail stores, clinics, tutoring centers, factories, and front-desk areas.

This article walks through the fundamentals of face recognition setup — including how to adjust parameters and how to enroll employee face data. You can also follow along with our video tutorial.

What You Need Before Getting Started

Before you begin, prepare the following:

A phone or tablet that can be mounted in a fixed position
The TimeClock app installed
Employee profiles already created
A stable lighting environment at the mounting location

Mount the device somewhere employees naturally pass when clocking in and out — near an entrance, beside the front desk, at the office door, or close to a break room.

The camera height should be roughly at face level. Avoid mounting it too low or too high, as extreme angles reduce recognition accuracy.

What Are Face Recognition Parameters?

Face recognition parameters control how the system decides whether a face matches an employee's profile.

The settings you're most likely to adjust include: recognition threshold, liveness detection, recognition distance, recognition interval, and duplicate check-in prevention.

Label names may vary slightly between app versions, but the underlying logic is the same.

Setting the Recognition Threshold

The recognition threshold determines how strict the system is when comparing a live face to stored face data.

A higher threshold makes the system stricter and more secure, but some employees may need a few attempts to get through.
A lower threshold makes recognition easier to pass, but can reduce accuracy.

We recommend starting with the system default and letting employees use it for a while. If recognition fails too often, adjust gradually.

Avoid setting the threshold too low right away — it can compromise reliability.

Enable Liveness Detection

If the app includes a liveness detection feature, we recommend turning it on.

Liveness detection helps prevent someone from clocking in using a photo, a screenshot, or other non-real-person methods.

When enabled, employees may need to follow on-screen prompts — such as facing the camera directly or holding still — so the system can confirm they are physically present.

If preventing proxy clock-ins is a priority for your business, liveness detection is an important setting to enable.

Recognition Interval and Duplicate Check-In Prevention

If an employee lingers in front of the camera or passes it multiple times in quick succession, the system may record the check-in more than once.

To prevent multiple entries within a short window, you can configure a recognition interval or enable duplicate check-in prevention.

For example, the same employee might only be allowed one successful check-in within a set number of minutes.

This keeps your records clean and reduces noise in the dashboard.

How to Enroll Employee Face Data

Before face recognition can work, you need to enroll face data for each employee.

Here's the recommended process:

Create the employee profile in the dashboard or app.
Open the face data settings page for that employee.
Ask the employee to stand in front of the device with their full face visible in the frame.
Follow the on-screen prompts to capture or record face data.
Save the data and immediately run a test check-in.

During enrollment, ask employees to maintain a natural expression. Avoid having their face obscured by a mask, cap brim, sunglasses, or strong shadows.

If an employee regularly wears glasses at work, enroll them while wearing glasses so the system matches their real-world appearance.

Things to Watch Out for During Enrollment

The quality of the face data directly affects recognition success rates down the line.

Keep these points in mind:

Make sure lighting is stable — avoid backlighting.
Don't mount the camera too low or too high.
The employee's full face must be visible in the frame.
Avoid having multiple people in the frame at the same time.
Always run a test check-in immediately after enrollment.

If a particular employee has inconsistent recognition results, re-enroll their face data or adjust the device position and lighting conditions.

Closing Thoughts

Face recognition check-in is a great fit for companies that don't want to manage cards, or where employees often forget to bring them.

Start with the default parameters and adjust the recognition threshold, liveness detection, and duplicate prevention settings based on how things go in practice.

With clear face data, a stable device placement, and adequate lighting, employees can clock in and out in the most intuitive way possible — hands free.