“How can one create a real-time face detection system from live video, ensuring accuracy and efficiency, using modern computer vision techniques and frameworks?”