Finds moving regions and frames them with CCTV-style boxes, brackets and labels. Labels are decorative, not real recognition.
