Train the AI vision model by spotting moving subjects and drawing precise bounding boxes around them.