Skip to main content

Export a Clea Studio AI workflow to Arduino VENTUNO Q

Overview​

This procedure creates a face-detection workflow in Clea Studio AI, exports it for Arduino, imports the generated package into Arduino App Lab, and runs the resulting application on Arduino VENTUNO Q.

The example project is named Face Detection on Camera. Studio AI provides the visual AI workflow, while App Lab provides the Arduino application environment used on VENTUNO Q.

What you need in advance​

Prepare:

  • access to Clea Studio AI;
  • Arduino VENTUNO Q;
  • Arduino App Lab installed on the development computer;
  • VENTUNO Q connected to App Lab;
  • a camera supported by the final App Lab application;
  • the input video asset used for the Studio AI preview.

1. Create the project in Studio AI​

Open Clea Studio AI.

Create a new project and set the project name to:

Face Detection on Camera

Open the project editor.

The editor is divided into the following block groups in the left panel:

  • Input
  • Model
  • Output

Build the workflow with these blocks:

  • StaticVideoInput
  • Face Detection HMI
  • Image logger

Place the blocks on the canvas.

Connect the image output from StaticVideoInput to the input of Face Detection HMI.

Connect the detection output from Face Detection HMI to Image logger.

The resulting flow is:

StaticVideoInput > Face Detection HMI > Image logger

2. Configure the Studio AI input​

Select the StaticVideoInput block.

Configure the input used for the project preview.

The reference configuration uses:

Source: video_face_detection.mp4
Drop policy: target-fps
Drop policy value: 15
Frame width: 320
Frame height: 240

Use the source file available in your Studio AI workspace if its filename differs.

Open the project settings panel and confirm the project directory or workspace selected for the face-detection component.

Save the project.

3. Run the Studio AI preview​

Start the project preview from the project editor.

Watch the output from Image logger.

The preview should show the input frames with the face-detection result generated by the model.

Check that:

  • the input video is read correctly;
  • the face detector receives frames;
  • detections are visible in the output;
  • the project remains in a running state.

Keep this preview as the reference for the VENTUNO Q run.

4. Export the project for Arduino​

From the Studio AI project editor, select:

Deploy to edge

The deployment dialog opens.

For Target device, select:

Arduino

Click:

Download

Studio AI generates the Arduino package and downloads it to the development computer.

Keep the ZIP file generated by Studio AI. In the reference workflow, the downloaded package is named after the project, for example:

face-detection-on-camera.zip

5. Import the package into Arduino App Lab​

Open Arduino App Lab and connect to VENTUNO Q.

Go to:

My Apps

In the upper-right corner, click:

Create new app +

Select:

Import App

The Import an App dialog opens.

Click:

Import from computer

Select the ZIP downloaded from Studio AI and confirm the import.

After App Lab completes the operation, the generated application appears in My Apps. For this workflow, the application is shown as:

Face Detection on Camera SECO

Open the application.

6. Inspect the generated App Lab project​

In the App Lab editor, review the application structure before running it.

Open the file tree and check the generated application files and Brick content.

The Studio AI component is carried into the App Lab project as part of the generated application package. App Lab then manages the application runtime on VENTUNO Q.

Open:

Files > python > main.py

Check that the generated application entry point is present.

Review the Bricks section as well. The face-detection component generated from Studio AI should be available to the application.

No manual recreation of the Studio AI workflow is required at this stage.

7. Run the application on VENTUNO Q​

Connect the camera required by the application.

Click:

Run

in the upper-right corner of App Lab.

If another application is already active, App Lab can display a Swap Running App dialog. Confirm the replacement to start the imported application.

Wait for the runtime to initialize.

Check the application output and compare it with the Studio AI preview.

Verify:

  • the application starts successfully;
  • the face-detection component initializes;
  • the input source is available;
  • faces are detected;
  • the output remains stable over multiple frames.

If the VENTUNO Q application uses a live camera while the Studio AI preview used a video file, compare the detection behavior and model output. Capture-related differences can be expected because the two inputs are different.

8. Check the application from App CLI​

For an additional runtime check, open the board shell from App Lab or through your usual terminal connection.

List the applications:

arduino-app-cli app list

Find the App ID associated with Face Detection on Camera SECO.

Inspect the logs:

arduino-app-cli app logs <APP_ID>

This is useful when the App Lab UI reports a startup problem or when the AI component initializes but no output reaches the application interface.

Result​

The Studio AI workflow is now available as an Arduino App Lab application and has been executed on VENTUNO Q.

The resulting App Lab project can also be exported with:

App name menu > Export App > Export .zip

That package can then follow the deployment procedure described in the Deploy an Arduino App Lab application to SOM-SMARC-Dragonwing-IQ8 guide.

Clea Studio AI
Arduino App Lab examples
Arduino App specification