User Guide: Adding and Configuring an Application

This guide explains how to open an application from App Stack, attach devices, configure ROI, set meta parameters, save the configuration, and activate it.

1. Select the Application

Open App Stack and search for the application you want to configure. The example below shows the application list filtered for a vehicle application.

1. Navigate to App Stack from the top menu.
2. Use the search bar or category list to find the required application.
3. Open the application and start the configuration flow.

2. Configure Devices

The first step in the configuration flow is Device Configuration. Use this page to attach the cameras that should run the selected application.

1. Click Add Device.
2. Select the required camera or camera group from the device list.
3. Click Add to attach the selected devices.

After devices are added, verify that the selected feeds are visible and mapped correctly before moving to ROI configuration.

3. Configure ROI

ROI configuration defines where the application should observe activity in the camera feed. The available ROI type depends on the selected application.

1. Select the device feed.
2. Click Add ROI.
3. Draw the ROI on the video frame.
4. Enter the ROI details and save it.

Use the appropriate ROI type for the application:

The saved ROI is then available for selection in the meta-parameter flow.

Review the configured ROI and make sure the direction, color, and region placement are correct before continuing.

4. Configure Meta Parameters

The Meta Parameters step controls how the application pipeline runs. It connects the device, detection logic, ROI, event generation, and alert configuration.

4.1 Device Node

Open the Device node and select the device or devices that should be used in this configuration.

Common device-level settings can include:

• Device selection
• Direction or movement settings
• Application-specific options
• Camera-wise processing options

4.2 Detection Node

Open the application detection node and configure model-related options such as confidence threshold, run method, model type, or selected classes.

Recommended guidance:

• Keep the default confidence value unless the deployment needs tuning.
• Increase confidence to reduce false positives.
• Lower confidence carefully only when detections are being missed.
• Confirm model and class selections with the CV or deployment team when needed.

4.3 ROI Node

Open the ROI node and select the ROI created in the previous step.

Make sure the selected ROI belongs to the correct device. For line-based applications, also verify the direction setting before saving.

4.4 Event Node

The Event node controls what information is stored when the application generates an event.

Depending on the application, event settings may include class, confidence, bounding box, ROI, image evidence, or video evidence.

4.5 Alert Node

The Alert node controls whether event alerts are generated and how they are routed.

Configure alert behavior based on the use case. Enable alerts only when the event needs to notify users, operators, or external systems.

5. Save the Configuration

After all nodes are configured, save the application configuration.

Before saving, confirm:

• The correct device is selected.
• The correct ROI is selected.
• Required meta parameters are configured.
• Hardware type is selected as CPU or GPU based on deployment.
• Batch size is set as recommended for the server.

6. Activate or Update the Configuration

After saving, use the configuration controls to activate, update, or reset the setup.

• Turn the configuration On to start processing.
• Use Update after making changes to an existing configuration.
• Use Reset only when you want to discard unsaved changes.

7. Verify Events

Once the configuration is active, verify that events are being generated as expected.

Check the generated events for correct device name, ROI mapping, event type, timestamp, and evidence. If the results are not accurate, revisit the ROI and confidence settings.

Key Benefits

• Reusable flow: The same configuration steps apply across many App Stack applications.
• Configurable detection: Each node can be tuned for the deployment environment.
• Controlled activation: Applications can be saved, reviewed, and activated only when ready.