To provide an apartment building interphone system which can deduce a suspicious person from a video image taken by means of a camera and can detect intrusion of the suspicious person through an automatic door.

A hallway intercom camera 10 for picking the image of a range including a visitor and the vicinity of an auto door, a hallway intercom display section 11 for outputting a video image picked up by means of the hallway intercom camera 10, an image analysis section 16 having an algorithm for recognizing and tracking a person in the video image thus picked up, and a touch panel 12 for specifying a specific person admitted to enter a building are provided in a collective entrance unit 1. An area for watching intrusion of a person is set near an auto door inlet in the imaging range and when the image analysis section 16 tracks a specific person specified by the touch panel 12 and an unspecified person and detects the intrusion of the unspecified person into the watch area, a hallway intercom control section 19 triggers alarming operation of a speaker in a speech communication section 13.