no points can be
recognized. It can be observed that for most of the frames,
video from person 1 has given more recognized points than
person 2 and person 3. Thus a threshold value for recognized
points or features can be chosen in order to distinguish the face
of an intruder from a face of a house member.