feat(selfchkout): add UIs and scripts for visual interaction & easy demo

[mjdivankoller]

The update focuses on the Automated Self-checkout reference implementation.

• It incorporates a new Gradio-based UI that provides interactivity and additional options to view, process, and review the video (See the Jupyter Notebook), detections/tracking, and log messages.
  • It can choose a different video for previewing (full or small screen) and processing it.
  • The processed video is shown when the batch of frames is released (progressive).
  • It can save the final processed video for future reference.
  • A Log table allows the user to filter outputs dynamically while the video is under processing or once processed.
  • A bar plot shows the detected object classes and their frequency.
  • The purchased items are shown as an independent table to contrast them with log messages and detected objects.
• It adds a set of scripts to make it easier to test the demo. For example, you can install, run, or clean the environment and run the demo visually directly (See the 'Scripts' section).