Image ghost removal

• US 9,218,669 B1
• Filed: 01/18/2012
• Issued: 12/22/2015
• Est. Priority Date: 01/20/2011
• Status: Active Grant

• Current Assignee: Cognyte Technologies Israel Ltd. (Cognyte Software Ltd.)
• Original Assignee: Verint Systems Limited (Verint Systems Incorporated)
• Inventors: Golan, Oren, Horovitz, Itshak
• Primary Examiner(s): Liu, Li
• Application Number:
  US13/352,642
• Time in Patent Office: 1,434 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 7/254 involving subtraction of images

Abstract

A method for image ghost removal is provided. The method for image ghost removal includes receiving an image and a background model related to the image, and generating a foreground mask based on the image and the background model. The method also includes identifying image ghosts within the foreground mask, and updating the background model to eliminate the image ghosts.

Claims

1. A method for image ghost removal comprising:

receiving an image and a first background model related to the image;

generating a foreground mask based only on the image and the first background model;

identifying image ghosts within the foreground mask; and

updating the first background model to eliminate the image ghosts without comparing the image to a second background model, wherein the identifying image ghosts entails a determination of an array of counts for each pixel in the image containing a number of seconds that the pixel has been continuously flagged as part of the ghost image in the foreground mask,

and wherein the identifying image ghosts further entails measuring an amount of motion within the image ghosts to ensure that the image ghosts are not a foreground element.

2. The method of claim 1, wherein identifying image ghosts comprises determining pixels within the foreground mask that contain a background element incorrectly identified as a foreground element.

3. The method of claim 1, wherein generating the foreground mask comprises comparing the image to the first background model and including only pixels that are different from the background model in the foreground mask.

4. The method of claim 3, wherein identifying image ghosts comprises:

determining a length of time that each pixel in the foreground mask has been different from a corresponding pixel in the first background image; and

identifying a pixel as an image ghost pixel when the length of time for that pixel exceeds a criteria.

5. The method of claim 4, wherein updating the first background model comprises overwriting pixels corresponding to the image ghost in the background model with the corresponding pixels in the image.

6. The method of claim 5, further comprising:

removing the image ghost from the foreground mask.

7. The method of claim 6, further comprising:

storing the updated background model in a memory.

8. A computer system configured for image ghost removal, the computer system comprising:

a memory configured to store images and background models; and

a processor coupled to the memory configured to;

receive an image and a first background model related to the image;

generate a foreground mask based only on the image and the first background model;

identify image ghosts within the foreground mask; and

update the first background model to eliminate the image ghosts without comparing

the image to a second background model, wherein the identifying image ghosts entails a determination of an array of counts for each pixel in the image containing a number of seconds that the pixel has been continuously flagged as part of the ghost image in the foreground mask,

and wherein the identifying image ghosts further entails measuring an amount of motion within the image ghosts to ensure that the image ghosts are not a foreground element.

9. The computer system of claim 8, wherein identifying image ghosts comprises determining pixels within the foreground mask that contain a background element incorrectly identified as a foreground element.

10. The computer system of claim 8, wherein generating the foreground mask comprises comparing the image to the first background model and including only pixels that are different from the first background model in the foreground mask.

11. The computer system of claim 10, wherein identifying image ghosts comprises:

determining a length of time that each pixel in the foreground mask has been different from a corresponding pixel in the first background image; and

identifying a pixel as an image ghost pixel when the length of time for that pixel exceeds a criteria.

12. The computer system of claim 11, wherein updating the first background model comprises overwriting pixels corresponding to the image ghost in the first background model with the corresponding pixels in the image.

13. The computer system of claim 12, wherein the processor is further configured to:

remove the image ghost from the foreground mask.

14. The computer system of claim 13, wherein the processor is further configured to:

store the updated background model in the memory.

15. A non-transitory computer readable medium having stored thereon program instructions for image ghost removal, that when executed by a computer system, direct the computer system to:

receive an image and a first background model related to the image;

generate a foreground mask based only on the image and the first background model;

identify image ghosts within the foreground mask; and

update the first background model to eliminate the image ghosts without comparing the image to a second background model

wherein the identifying image ghosts entails a determination of an array of counts for each pixel in the image containing a number of seconds that the pixel has been continuously flagged as part of the ghost image in the foreground mask,

and wherein the identifying image ghosts further entails measuring an amount of motion within the image ghost to ensure that the image ghosts are not a foreground element.

16. The non-transitory computer readable medium of claim 15, wherein identifying image ghosts comprises determining pixels within the foreground mask that contain a background element incorrectly identified as a foreground element.

17. The non-transitory computer readable medium of claim 15, wherein generating the foreground mask comprises comparing the image to the first background model and including only pixels that are different from the first background model in the foreground mask.

18. The non-transitory computer readable medium of claim 17, wherein identifying image ghosts comprises:

determining a length of time that each pixel in the foreground mask has been different from a corresponding pixel in the first background image; and

identifying a pixel as an image ghost pixel when the length of time for that pixel exceeds a criteria.

19. The non-transitory computer readable medium of claim 18, wherein updating the first background model comprises overwriting pixels corresponding to the image ghost in the background model with the corresponding pixels in the image.

20. The non-transitory computer readable medium of claim 19, wherein the instructions further direct the computer system to:

remove the image ghost from the foreground mask.

21. The non-transitory computer readable medium of claim 20, wherein the instructions further direct the computer system to:

store the updated background model in the memory.