为每次检测生成稳定的面部id

Question

我使用以下代码从早到晚统计面孔。

 while True:
        ret, frame = video_capture.read()
        gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
        detections = faceCascade.detectMultiScale(gray, 1.15, 5)

        current_viewers = len(detections)
        if current_viewers > last_total_current_viewers:
            people_count += current_viewers - last_total_current_viewers
        last_total_current_viewers = current_viewers

        for i in range(len(detections)):
            face_i = detections[i]
            x, y, w, h = face_i
            cv2.rectangle(frame, (x, y),
                          (x + w, y + h),
                          (222,222,222), 2)
            cv2.putText(frame, str(people_count - i ), (x, y - 10), font, 0.5, (255, 255, 255), 1, cv2.LINE_AA)

以上代码工作正常，但不稳定。如果用户摇头或用户快速移动，则计数迅速增加。如何跟踪每个并保持计数不变？