I need a straightforward Python program that runs smoothly on a typical Linux workstation (Ubuntu 20.04 on my end) and leverages OpenCV together with the face_recognition library. The goal is simple: I point the script at one or more PNG images containing the people I want to track, start the webcam, and instantly see bounding boxes and their names appear around any matching faces in the live video stream. Core flow I expect • Load reference PNGs, generate encodings, and store them in memory. • Open the default webcam, read frames in real time, detect every face, compare encodings, and label any match. • Draw a rectangle and overlay the matching person’s name. Unrecognised faces stay unlabelled. • A single command-line flag (or a config section) lets me choose between only one reference image or a whole folder of them so I can switch between “single” and “multiple” modes without rewriting code. • Pressing “q” cleanly breaks the loop, releases the camera, and closes all OpenCV windows. Deliverable A single, well-commented .py file plus a short README that lists the required pip installs and demonstrates the one-versus-many switch. Everything should run with Python 3.9, OpenCV-Python, and face_recognition installed from PyPI. Timeline I’d like the finished script, fully tested, within the coming month, so there’s time for a quick tweak round if needed. I’ll test by pointing the camera at myself (one of the provided PNGs) and at others who are not listed; correct identification and clean exit will mark completion.