Methods and apparatus to count people in images

• US 8,761,442 B2
• Filed: 03/29/2012
• Issued: 06/24/2014
• Est. Priority Date: 03/29/2012
• Status: Active Grant

• Current Assignee: The Nielsen Company LLC (MOOD Media Corporation)
• Original Assignee: The Nielsen Company LLC (MOOD Media Corporation)
• Inventors: Srinivasan, Venugopal, Soundararajan, Padmanabhan
• Primary Examiner(s): CHANG, JON CARLTON
• Application Number:
  US13/434,319
• Publication Number:
  US 20130259299A1
• Time in Patent Office: 817 Days
• Field of Search: None
• US Class Current: 382/103
• CPC Class Codes: G06V 20/53 Recognition of crowd images, e.g. recognition of crowd congestion
• CPC Class Codes: G06V 40/173 face re-identification, e.g. recognising unknown faces across different face tracks

Abstract

Methods and apparatus to count people in images are disclosed. An example method includes, in response to determining that a first person indication of a first frame obtained via a first image sensor is redundant to a second person indication of a second frame obtained via a second image sensor, storing a first coordinate of the first person indication in connection with a second coordinate of the second person indication in a database; and in response to detecting a third person indication in a third frame obtained via the first image sensor, querying the database with a third coordinate of the third person indication to determine whether the third coordinate matches the first coordinate.

Claims

1. A method, comprising:

in response to determining that a first person indication of a first frame obtained via a first image sensor is redundant to a second person indication of a second frame obtained via a second image sensor, storing a first coordinate of the first person indication in connection with a second coordinate of the second person indication in a database; and

in response to detecting a third person indication in a third frame obtained via the first image sensor, querying the database with a third coordinate of the third person indication to determine whether the third coordinate matches the first coordinate.

2. A method as defined in claim 1, further comprising, when the third coordinate matches the first coordinate, determining whether a fourth frame includes a fourth person indication at the second coordinate.

3. A method as defined in claim 2, further comprising marking the third person indication as redundant when the fourth frame includes the fourth person indication at the second coordinate.

4. A method as defined in claim 1, wherein a first frame pair formed by the first and second frames is captured at a first time, and wherein a second frame pair formed by the third frame and a fourth frame is captured at a second time later than the first time.

5. A method as defined in claim 1, wherein the first person indication is determined to be redundant to the second person indication by comparing first image data of the first person indication to second image data of the second person indication.

6. A method as defined in claim 5, wherein the comparison of the first image data and the second image data is triggered by a detection of the first person indication in an overlap region associated with fields of view of the first and second image sensors.

7. A method as defined in claim 1, further comprising detecting human faces to generate the first and second person indications.

8. A tangible machine readable storage medium comprising instructions that, when executed, cause a machine to at least:

in response to determining that a first person indication of a first frame obtained via a first image sensor is redundant to a second person indication of a second frame obtained via a second image sensor, store a first coordinate of the first person indication in connection with a second coordinate of the second person indication in a database;

in response to detecting a third person indication in a third frame obtained via the first image sensor, query the database with a third coordinate of the third person indication to determine whether the third coordinate matches the first coordinate.

9. A machine readable medium as defined in claim 8, wherein the instructions, when executed, cause the machine to, when the third coordinate matches the first coordinate, determine whether a fourth frame includes a fourth person indication at the second coordinate.

10. A machine readable medium as defined in claim 9, wherein the instructions, when executed, cause the machine to mark the third person indication as redundant in the second frame pair when the fourth frame includes the fourth person indication at the second coordinate.

11. A machine readable medium as defined in claim 8, wherein a first frame pair formed by the first and second frames is captured at a first time, and wherein a second frame pair formed by the third frame and a fourth frame is captured at a second time later than the first time.

12. A machine readable medium as defined in claim 8, wherein the first person indication is determined to be redundant to the second person indication by comparing first image data of the first person indication to second image data of the second person indication.

13. A machine readable medium as defined in claim 12, wherein the comparison of the first image data and the second image data is triggered by a detection of the first person indication in an overlap region associated with fields of view of the first and second image sensors.

14. A machine readable medium as defined in claim 8, wherein the instructions, when executed, cause the machine to detect human faces to generate the first and second person indications.

15. An apparatus, comprising:

a database to store a first coordinate of a first person indication of a first frame in connection with a second coordinate of a second person indication of a second frame, the first person indication obtained via a first image sensor being redundant to the second person indication obtained via a second image sensor; and

a searcher to, in response to detecting a third person indication in a third frame obtained via the first image sensor, query the database with a third coordinate of the third person indication to determine whether the third coordinate matches the first coordinate.

16. An apparatus as defined in claim 15, further comprising an analyzer to, when the third coordinate matches the first coordinate, determine whether a fourth frame includes a fourth person indication at the second coordinate.

17. An apparatus as defined in claim 16, wherein the analyzer is to identify the third person indication as redundant in the second frame pair when the fourth frame includes the fourth person indication at the second coordinate.

18. An apparatus as defined in claim 15, wherein a first frame pair formed by the first and second frames is captured at a first time, and wherein a second frame pair formed by the third frame and a fourth frame is captured at a second time later than the first time.

19. An apparatus as defined in claim 15, further comprising a comparator to determine that the first person indication is redundant to the second person indication by comparing first image data of the first person indication to second image data of the second person indication.

20. An apparatus as defined in claim 19, wherein the comparison of the first image data and the second image data is triggered by a detection of the first person indication in an overlap region associated with fields of view of the first and second image sensors.