Social network construction based on data association

• US 8,284,990 B2
• Filed: 02/11/2009
• Issued: 10/09/2012
• Est. Priority Date: 05/21/2008
• Status: Active Grant

• Current Assignee: Honeywell International Inc.
• Original Assignee: Honeywell International Inc.
• Inventors: Ma, Yunqian, Schloegel, Kirk, Miezianko, Roland, Cisar, Petr
• Primary Examiner(s): BHATNAGAR, ANAND P
• Application Number:
  US12/369,692
• Publication Number:
  US 20090292549A1
• Time in Patent Office: 1,336 Days
• Field of Search: 382/100, 382/103, 382/181, 382/190, 382/195, 348/135, 348/143, 348169-172
• US Class Current: 382/103
• CPC Class Codes: G06Q 50/01 Social networking
• CPC Class Codes: G06V 20/30 in albums, collections or shared content, e.g. social network photos or video
• CPC Class Codes: G06V 20/44 Event detection
• CPC Class Codes: G06V 20/52 Surveillance or monitoring of activities, e.g. for recognising suspicious objects recognising microscopic objects G06V20/69
• CPC Class Codes: G06V 40/20 Movements or behaviour, e.g. gesture recognition recognition of facial expressions G06V40/16
• CPC Class Codes: H04N 7/18 Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast

Abstract

A system for social network construction. Video analytics and association may be used to develop a social network. Also, social groups may be developed from temporal proximity of persons. In the case of several social networks, they may be collapsed into one network or a weighted graph that mining algorithms can handle.

Claims

1. A system for discovering a social group, comprising:

a plurality of cameras for providing video data from various locations; and

a video analytics mechanism connected to the plurality of cameras; and

wherein;

the video analytics mechanism is for recognizing a person of interest in video data, identifying people in temporal proximity to the person of interest from the video data, and for identifying a temporal pattern of people in temporal proximity to the person of interest;

the temporal pattern of people in temporal proximity to the person of interest is identified with temporal sequence matching; and

the temporal pattern reveals a social group.

2. The system of claim 1, wherein the person of interest is pre-selected for surveillance.

3. The system of claim 2, wherein:

being pre-selected is manual or automatic;

manual involves another person effecting the person of interest to be pre-selected; and

automatic involves a detection module effecting the person of interest to be pre-selected.

4. The system of claim 2, wherein temporal proximity is a time gap no longer than a specified period of time.

5. The system of claim 2, wherein people in temporal proximity to the person of interest are identified across at least some of the plurality of cameras.

6. The system of claim 2, wherein the person of interest constitutes more than one person.

7. The system of claim 2, wherein the video analytics mechanism utilizes face recognition for identifying a person of interest in video data and for identifying people in temporal proximity to the person of interest from the video data.

8. The system of claim 2, wherein the temporal sequence matching includes:

detecting two or more temporal patterns of the persons in one or more of the plurality of cameras;

comparing the two or more temporal patterns to obtain a common temporal pattern of the persons in the one or more of the plurality of cameras; and

the common temporal pattern reveals a social group.

9. The system of claim 1, wherein the temporal sequence matching includes:

identifying a temporal sequence of people in temporal proximity to the person of interest at a first time; and

identifying the temporal sequence of people in temporal proximity to the person of interest at a second time.

10. The system of claim 9, wherein the temporal sequence matching is performed with video data from a single camera, where the video data is captured at the first time and at the second time with the single camera.

11. The system of claim 1, wherein the temporal sequence matching includes:

identifying a temporal sequence of people in temporal proximity to the person of interest with a first camera; and

identifying the temporal sequence of people in temporal proximity to the person of interest with a second camera.

12. The system of claim 1, wherein the people in temporal proximity of the person of interest and the person of interest do not appear in a field of view of any of the plurality of cameras at the same time.

13. A system for discovering a social group, comprising:

a plurality of cameras for providing video data from various locations; and

a video analytics mechanism connected to the plurality of cameras; and

wherein;

the video analytics mechanism is for recognizing a person of interest in video data, identifying people in temporal proximity to the person of interest from the video data, and for identifying a temporal pattern of people in temporal proximity to the person of interest;

temporal proximity comprises;

one or more people present at a location at a first time;

a person of interest present at the location at a second time;

one or more people present at the location at a third time;

the location is a field of view of each of two or more cameras; and

the fields of view of the two or more cameras do not necessarily overlap; and

the temporal pattern reveals a social group.

14. The system of claim 13, wherein the person of interest is pre-selected for surveillance.

15. The system of claim 13, wherein:

the temporal pattern of people in temporal proximity to the person of interest is identified through temporal sequence matching; and

the temporal sequence matching includes matching a first temporal proximity with a second temporal proximity.

16. A system for discovering a social group, comprising:

a camera system for providing video data from one or more locations; and

a video analytics mechanism connected to the camera system; and

wherein;

the video analytics mechanism is for identifying one or more people in temporal proximity to a person of interest from the video data, and for identifying a temporal pattern of people in temporal proximity to the person of interest;

the temporal pattern of people in temporal proximity to the person of interest is identified through identifying a temporal sequence of people in temporal proximity to the person of interest at a first time and identifying the temporal sequence of people in temporal proximity to the person of interest at a second time;

the first time is discrete from the second time; and

the temporal pattern reveals a social group.

17. The system of claim 16, wherein the temporal pattern is determined using a single camera of the camera system.

18. The system of claim 16, wherein:

the temporal sequence of people in temporal proximity to the person of interest at the first time is identified with a first camera; and

the temporal sequence of people in temporal proximity to the person of interest at the second time is identified with a second camera.

19. The system of claim 16, wherein the person of interest is automatically pre-selected with a detection module selecting the person of interest.

20. The system of claim 16, wherein a temporal sequence is a sequence including the time at which one or more people in temporal proximity to the person of interest are present at a location with respect to the time at which the person of interest is present at the location.