AI Models Creation and Import

In ScreenManager, you can import an AI model—a simple mapping of the object classes your camera can detect—formatted as:

<classId>:<label> (for example: 10:apple)

This shared list is important because it ensures the camera and the signage player “speak the same language.” The class IDs and labels you import into ScreenManager must match what your AI camera actually outputs.

1. Camera detects an object Example: the camera detects an apple and returns the class ID 10.
2. Camera sends the detected class ID to the player The player receives something like: 10.
3. Player resolves the class ID using your imported AI model ScreenManager looks up 10 in the AI model mapping and translates it to apple.
4. Playlist rules decide what to play In your playlist, you configure rules such as: “If apple is detected, play Content X.”
5. Player switches content automatically The signage player immediately switches to Content X based on the detection.

This approach lets you drive playlist behavior using real-time object detection—without hardcoding logic—because the player can reliably interpret camera detections through the imported AI model mapping.

1. Once you are logged into your ScreenManager account, navigate to the organization page.
2. Click on “Create New Model”.

3. Choose a name for your model (Fruits in our example) and insert labels of the objects you want to detect.
For example add the below to allow AI to track fruits, usually this data are provided by the creator of the AI model on the AI sensor.

1:apple

2:banana

3:orange

4. Click on “Create New AI Model”

After the model is created, you can view the AI model's labels by clicking on “View labels” button.