Training Faces, Plates, and Custom Classes

Enroll faces and plates, or train the custom analyses, so the AI can tell who someone is or what state something is in. Enrollment and training also work while running.

Enrolling Faces

Open Enroll from a face condition row or the analysis settings.
Click Add a name and enter the person's name.
With that name selected, have the person face the camera and click Enroll. Both eyes must be visible. Side profiles are not enrolled.
Repeat five or more times, varying the angle and expression slightly.

The face enrollment panel — The enrollment panel. Create the name first, keep it selected, and enroll from the camera.

When someone is recognized as another person, and blocking photo spoofing

If a person is recognized as someone else, raise the face Threshold and enroll additional photos of both people.

Where recognition grants access, turn on Anti-Spoofing in the analysis settings. It rejects faces presented on photos or phone screens.

Registering Plates

Open Register from a plate condition row.
Enter the number. Enter an alias too if you like. An alias can also be a target value, so a condition can match "delivery van" instead of the number.
Set Region in the analysis settings to the installation country. With the wrong region, reading accuracy stays poor regardless of the confidence setting.

The Registered Plates panel with plate and alias fields — The plate registry. Register plates with an optional alias, or paste a list at once with Bulk add.

Registering many plates at once

Paste a list into Bulk add, one per line, as the plate alone or in plate,alias format.

Training Custom Classification

Open Train from the condition row or the analysis settings.
Create at least two classes and name them.A classifier with one class has nothing to decide between, so two is the minimum.
Select a class, set up the scene, and capture. Repeat for each class.
Click Train and wait for completion.Samples added after the last training are marked as untrained.

The training panel. Captured samples for the selected class on the left, the class list on the right.

Training Custom Anomaly Detection

Anomaly detection learns only the normal state and reports anything that deviates from it. No fault samples are needed, so it also covers problems that have not happened yet.

Open Add normals.
Capture the scene several times while it is in a normal state.
Click Train.When training finishes, the Normal score distribution appears.

Caution Anomaly detection suits places where camera angle, distance, and lighting are fixed, such as an inspection station. Where the scene keeps changing, such as a corridor, it produces continuous false alarms.

Interpreting the score chart after training

Training produces a Normal score distribution. Read it left to right: farther right is closer to a fault, and anything right of the red line is judged faulty. Training holds back about 20 percent of the registered samples to set that line, which is why the displayed count is smaller than the number registered.

The normal score distribution chart — Each dot is one held-out normal sample. Training succeeded if no dots fall right of the red line (the fault zone).

Distribution	Action
Dots grouped tightly left of the line	Training succeeded. No action needed.
Some dots past the line	Normal units in that condition may be judged faulty. Register more normals and train again.
Dots widely spread	Samples were captured under mixed lighting or angles. Register more under consistent conditions.