Methods and apparatus to count people in images

• US 9,092,675 B2
• Filed: 03/29/2012
• Issued: 07/28/2015
• 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
• Assistant Examiner(s): Mistry, ONeal R
• Application Number:
  US13/434,337
• Publication Number:
  US 20130259381A1
• Time in Patent Office: 1,216 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/30242 Counting objects in image
• CPC Class Codes: G06T 7/20 Analysis of motion motion estimation for coding, decoding, compressing or decompressing digital video signals H04N19/43, H04N19/51
• CPC Class Codes: G06T 7/70 Determining position or orientation of objects or cameras camera calibration G06T7/80
• 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 40/16 Human faces, e.g. facial parts, sketches or expressions
• CPC Class Codes: G06V 40/161 Detection; Localisation; Normalisation

Abstract

Methods and apparatus to count people in images are disclosed. An example method includes dividing a frame of image data into a plurality of segments; calculating first fluctuation factors for the respective segments; calculating a second fluctuation factor for the frame; and identifying ones of the segments having a first fluctuation factor greater than the second fluctuation factor as active segments.

Claims

1. A method, comprising:

dividing a first frame of image data into a plurality of segments;

calculating, with a processor, first fluctuation factors for respective ones of the segments, the first fluctuation factors respectively representative of changes in the respective segments from a second frame of image data to the first frame of image data;

calculating, with the processor, a second fluctuation factor for the first frame;

before performing object detection on the first frame;

identifying ones of the plurality of segments of the first frame having a first fluctuation factor greater than the second fluctuation factor as active segments for the first frame; and

linking, with the processor, adjacent ones of the active segments of the first frame to form a region of interest in the first frame; and

performing, with the processor, the object detection on the region of interest.

2. A method as defined in claim 1, wherein the adjacent ones of the active segments are adjacent within a margin of error.

3. A method as defined in claim 1, further comprising not performing the object detection on areas of the first frame outside the region of interest.

4. A method as defined in claim 1, wherein the linking of the adjacent ones of the active segments comprises forming a shape that includes the adjacent ones of the active segments.

5. A method as defined in claim 1, wherein the second fluctuation factor is calculated by selecting a set of pixels from the image data of the first frame and averaging fluctuation characteristics of the selected pixels.

6. A method as defined in claim 1, wherein the first and second fluctuation factors are root mean square values based on pixel intensities.

7. A method as defined in claim 1, wherein the second fluctuation factor is representative of a collective change across multiple ones of the segments over time.

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

divide a first frame of image data into a plurality of segments;

calculate first fluctuation factors for respective ones of the segments, the first fluctuation factors respectively indicative of fluctuations of the respective segments from a second frame of image data to the first frame;

calculate a second fluctuation factor for the first frame;

before performing object detection on the first frame;

identify ones of the plurality of segments of the first frame having a first fluctuation factor greater than the second fluctuation factor as active segments for the first frame; and

link a first one of the active segments of the first frame and a second one of the active segments of the first frame adjacent the first active segment to form a region of interest in the first frame; and

perform the object detection on the region of interest.

9. A tangible computer readable storage medium as defined in claim 8, wherein the first and second ones of the active segments are adjacent within a margin of error.

10. A tangible computer readable storage medium as defined in claim 8, wherein the instructions, when executed, cause the machine to perform the object detection on the region of interest and not perform the object detection on areas of the first frame not in the region of interest.

11. A tangible computer readable storage medium as defined in claim 8, wherein the linking of the first and second ones of the active segments comprises forming a shape that includes the first and second ones of the active segments.

12. A tangible computer readable storage medium as defined in claim 8, wherein the instructions, when executed, cause the machine to calculate the second fluctuation factor by selecting a set of pixels from the image data of the first frame and averaging fluctuation characteristics of the selected pixels.

13. A tangible computer readable storage medium as defined in claim 8, wherein the second fluctuation factor is representative of a collective change across multiple ones of the segments over time.

14. An apparatus, comprising:

a segmenter to divide first and second frames of image data into a plurality of segments;

a fluctuation calculator to respectively determine a first fluctuation factor for respective ones of the segments, the first fluctuation factors representing changes across the first frame and the second frame, the first and second frames captured at different times, the fluctuation calculator to determine a collective fluctuation factor for the first frame, the collective fluctuation factor indicative of a collective fluctuation of the first frame across the different times;

a comparator to, before performance of object detection on the first frame, mark ones of the segments of the first frame having a corresponding one of the first fluctuation factors greater than the collective fluctuation factor as active segments for the first frame; and

a linker to, before performance of the object detection on the first frame, combine adjacent ones of the active segments of the first frame to form a region of interest of the first frame; and

an object detector to perform the object detection on the region of interest of the first frame.

15. An apparatus as defined in claim 14, wherein the adjacent ones of the active segments are adjacent within a margin of error.

16. An apparatus as defined in claim 14, wherein the object detector comprises face detection logic, the face detection logic to operate on the region of interest but not on portions of the first frame outside the region of interest.

17. An apparatus as defined in claim 14, wherein the linker is to link the adjacent ones of the active segments by forming a shape that includes the adjacent ones of the segments.

18. An apparatus as defined in claim 14, wherein the fluctuation calculator is to calculate the collective fluctuation factor by selecting a random set of pixels from the image data of the first frame and averaging fluctuation characteristics of the randomly selected set of pixels.

19. An apparatus as defined in claim 14, wherein the first fluctuation factors and the collective fluctuation factor are root mean square values based on pixel intensities.