Kiosk mode for machine learning inference


I would like to run a Jetson device in ‘kiosk’ mode to perform object detection on images from a camera stream. The object detection inference is performed using a Python script, and currently results are displayed using OpenCV.

The device should boot up and start performing inference automatically, and it should output the results images (via HDMI port) and nothing else. A desktop should not be visible at any point.

Would anyone have any advice on the best/simplest way to achieve this?


You can write the inference task into a script and add it as follows to run automatically when boot.

The inference sample can be found in below GitHub:


This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.