Image processing methods

• US 7,929,729 B2
• Filed: 04/02/2007
• Issued: 04/19/2011
• Est. Priority Date: 04/02/2007
• Status: Active Grant

• Current Assignee: Industrial Technology Research Institute
• Original Assignee: Industrial Technology Research Institute
• Inventors: Cheng, Po-Feng, Huang, Yea-Shuan, Wang, Shih-Chun, Cheng, Hao-Ying
• Primary Examiner(s): Mariam; Daniel G
• Application Number:
  US11/695,573
• Publication Number:
  US 20080240500A1
• Time in Patent Office: 1,478 Days
• Field of Search: 382/103, 382/236, 382/107, 382/218, 382224-225, 382/190, 348/155
• US Class Current: 382/103
• CPC Class Codes: G06V 10/28 Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns

Abstract

A method of image processing, the method comprising receiving an image frame including a plurality of pixels, each of the plurality of pixels including an image information, conducting a first extraction based on the image information to identify foreground pixels related to a foreground object in the image frame and background pixels related to a background of the image frame, scanning the image frame in regions, identifying whether each of the regions includes a sufficient number of foreground pixels, identifying whether each of regions including a sufficient number of foreground pixels includes a foreground object, clustering regions including a foreground object into at least one group, each of the at least one group corresponding to a different foreground object in the image frame, and conducting a second extraction for each of at least one group to identify whether a foreground pixel in the each of the at least one group is to be converted to a background pixel.

Claims

1. A method of image processing, the method comprising:

receiving an image frame including a plurality of pixels, each of the plurality of pixels including an image information;

conducting a first extraction based on the image information to identify foreground pixels related to a foreground object in the image frame and background pixels related to a background of the image frame;

scanning the image frame in regions;

identifying whether each of the regions includes a sufficient number of foreground pixels;

identifying whether each of regions including a sufficient number of foreground pixels includes a foreground object;

clustering regions including a foreground object into at least one group, each of the at least one group corresponding to a different foreground object in the image frame; and

conducting a second extraction for each of at least one group to identify whether a foreground pixel in the each of the at least one group is to be converted to a background pixel,

wherein conducting the first extraction comprises applying a first background subtraction technique, and conducting the second extraction comprises applying a second background subtraction technique, and

wherein each of the at least one group includes a central portion and a boundary portion, further comprising;

using a third threshold for the central portion in conducting the second extraction; and

using a fourth threshold for the boundary portion in conducting the second extraction, the fourth threshold being greater than the third threshold.

2. The method of claim 1, wherein conducting the first extraction comprises applying one of a background subtraction technique or a temporal differencing technique.

3. The method of claim 2, wherein conducting the first extraction comprises using a background subtraction technique to identify foreground pixels related to a foreground object in the image frame and background pixels related to a background of the image frame.

4. The method of claim 1, wherein identifying whether each of the regions includes a sufficient number of foreground pixels comprises using an iterative operation.

5. The method of claim 1, wherein identifying whether each of regions including a sufficient number of foreground pixels includes a foreground object comprises using one of an Adaboost algorithm, a support method machine (SVM) algorithm or a neural network.

6. The method of claim 1, wherein conducting the second extraction comprises using one of a background subtraction technique or a temporal differencing technique.

9. The method of claim 1, wherein each of the at least one group includes a central portion and a lower portion, further comprising:

using a third threshold for the central portion in conducting the second extraction; and

using a fourth threshold for the boundary portion in conducting the second extraction, the fourth threshold being greater than the third threshold.

10. The method of claim 1, wherein each of the at least one group includes a central portion, a boundary portion and a lower portion, further comprising:

using a third threshold for the central portion in conducting the second extraction;

using a fourth threshold for the boundary portion in conducting the second extraction; and

using a fifth threshold for the lower portion in conducting the second extraction, the fifth threshold being greater than the fourth threshold, and the fourth threshold being greater than the third threshold.

11. The method of claim 1 further comprising assigning a score to each of the regions including a foreground object.

12. The method of claim 1 further comprising recording a position of each of the regions including foreground object.

13. The method of claim 1, wherein clustering regions including a foreground object into at least one group comprises:

identifying one of the regions including a foreground object as a seed region;

identifying whether an overlap between the seed region and one of the remaining regions is greater than a predetermined value; and

clustering the seed region and the one of the remaining region into a group if the overlap is greater than the predetermined value.

7. The method of claim , wherein the first background subtraction technique uses a first threshold and the second background subtraction technique uses a second threshold.

8. The method of claim 7, wherein the first threshold is greater than the second threshold.

14. A method of image processing, the method comprising:

receiving an image frame including a plurality of pixels;

conducting a first extraction to identify foreground pixels related to a foreground object in the image frame and background pixels related to a background of the image frame;

scanning the image frame in regions;

identifying whether each of the regions includes a sufficient number of foreground pixels;

identifying whether at least one of regions including a sufficient number of foreground pixel includes a foreground object;

identifying one of the at least one region including a foreground object as a seed region;

clustering the seed region and one of the at least one region into a group if an overlap between the seed region and the one of the at least one region is greater than a threshold; and

conducting a second extraction for the group to identify whether a foreground pixel in the group is to be converted to a background pixel in response to overlap between the seed region and the one of the at least one region being equal to or less than the threshold,

wherein conducting the first extraction comprises using a first background subtraction technique, and conducting the second extraction using a second background subtraction technique, and

wherein the group includes a central portion and a boundary portion, further comprising;

using a third threshold for the central portion in conducting the second extraction; and

using a fourth threshold for the boundary portion in conducting the second extraction, the fourth threshold being greater than the third threshold.

15. The method of claim 14 further comprising performing a logic OR operation for foreground pixels obtained by the first extraction and the second extraction.

16. The method of claim 14, wherein the first background subtraction technique uses a first threshold, and the second background subtraction technique uses a second threshold.

17. The method of claim 16, wherein the first threshold is greater than the second threshold.

18. The method of claim 14, wherein the group includes a central portion and a lower portion, further comprising:

using a third threshold for the central portion in conducting the second extraction; and

using a fourth threshold for the lower portion in conducting the second extraction, the fourth threshold being greater than the third threshold.

19. The method of claim 14, wherein the group includes a central portion, a boundary portion and a lower portion, further comprising:

using a third threshold for the central portion in conducting the second extraction;

using a fourth threshold for the boundary portion in conducting the second extraction; and

using a fifth threshold for the lower portion in conducting the second extraction, the fifth threshold being greater than the fourth threshold, and the fourth threshold being greater than the third threshold.

20. A method of image processing, the method comprising:

receiving an image frame including a plurality of pixels;

conducting a first background subtraction to identify foreground pixels related to a foreground object in the image frame and background pixels related to a background of the image frame;

identifying whether a region of the image frame includes a sufficient number of foreground pixels;

identifying whether a region including a sufficient number of foreground pixels includes a foreground object;

identifying whether in at least one region including a foreground object, a first region and a second region include the same foreground object;

conducting a second background subtraction for the group to identify whether a foreground pixel in the group is to be converted to a background pixel; and

updating one of the first and second background models with a first parameter, the first parameter satisfying;

B′

(P)=(1−

a) B(P)+a I(P), if a pixel P is identified as a background pixel and is not located in a region including at least a portion of an object,

wherein B′

(P) is an updated value of a P-th pixel in one of the first or second background models, I(P) is a value of the P-th pixel in the image, and B(P) is a value of the P-th pixel in one of the first or second background models.

21. The method of claim 20 further comprising establishing a first background model for the first background subtraction and a second background model from the second background subtraction.

22. The method of claim 21, wherein the first background model includes a first threshold and the second background model includes a second threshold, the first threshold being greater than the second threshold.

23. The method of claim 20 further comprising:

assigning a score to a region including a foreground object;

and recording a position of the region including the foreground object.

24. The method of claim 20 further comprising updating one of the first or second background models with a second parameter, the second parameter satisfying:

wherein B′

(P) is an updated value of a P-th pixel in one of the first or and second background models, I(P) is a value of the P-th pixel in the image, and B(P) is a value of the P-th pixel in one of the first or second background models.

25. The method of claim 20 further comprising updating one of the first or second background models with a third parameter, the third parameter satisfying:

B′

(P)=(1−

y)B(P)+y I(P), if a pixel P is identified as a foreground pixel,

wherein B′

(P) is an updated value of a P-th pixel in one of the first or second background models, I(P) is a value of the P-th pixel in the image, and B(P) is a value of the P-th pixel in one of the first or second background models.