Method for segmenting a digital image

• US 6,771,834 B1
• Filed: 07/02/1999
• Issued: 08/03/2004
• Est. Priority Date: 07/02/1999
• Status: Expired due to Term

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Martins, Fernando C. M., Hazra, Rajeeb
• Primary Examiner(s): LEE, TOMMY D
• Application Number:
  US09/347,481
• Time in Patent Office: 1,859 Days
• Field of Search: 382/173, 382/174, 382/178, 382/205, 382/256, 382/257, 382/260, 382/307, 382/308, 382/180, 358/464
• US Class Current: 382/257
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/20156 Automatic seed setting
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/194 involving foreground-background segmentation

Abstract

Briefly, in accordance with one embodiment of the invention, a method of segmenting an initial digital image includes the following. The initial digital image is processed to produce a first digital image with defined edges corresponding to the initial digital image and to produce a second digital image with at least two dominant contiguous regions corresponding to the initial digital image. Distinct non-overlapping regions of the first digital image formed by the defined edges are identified. The distinct non-overlapping regions of the first digital are combined based, at least in part, on a correspondence with the at least two dominant contiguous regions in the second digital image. Based, at least in part, on the remaining regions after combining the distinct non-overlapping regions of the first digital image, the initial digital image is segmented.

Claims

1. A method of segmenting an initial digital image comprising:

processing an initial digital image to produce a first digital image with defined edges corresponding to the initial digital image and a second digital image with at least two dominant contiguous regions corresponding to the initial digital image;

identifying distinct, non-overlapping regions of the first digital image formed by the defined edges;

combining the distinct, non-overlapping regions based, at least in part, on a correspondence with the at least two dominant contiguous regions of the second digital image; and

segmenting the initial digital image based, at least in part, on the remaining regions after combining the distinct, non-overlapping regions based, at least in part, on the correspondence with the at least two dominant contiguous regions.

2. The method of claim 1, wherein one segmentation comprises the background of the initial digital image.

3. The method of claim 2, wherein the remaining segmentation comprises the foreground of the initial digital image.

4. The method of claim 3, and further comprising segmenting additional digital images after said initial digital image.

5. The method of claim 4, wherein segmenting additional digital images after said initial digital image comprises:

determining the pixel locations in the background segmentation that have changed a measurable amount in pixel signal value and processing those pixel locations as in the foreground segmentation.

6. The method of claim 4, wherein segmenting additional digital images after said initial digital image comprises:

determining the pixel locations in the foreground segmentation that have not changed by an amount in pixel signal value that is specifically measurable and processing those pixel locations as in the background segmentation.

7. The method of claim 1, wherein identifying the distinct, non-overlapping regions comprises applying a region flood fill process.

8. The method of claim 7, wherein said region flood fill process comprises:

identifying a non-flooded pixel in the first image;

flood filling the distinct, non-overlapping region including the non-flooded pixel; and

continuing the immediately preceding identifying and flood filling operations until the first image contains no non-flooded pixels.

9. The method of claim 1, wherein processing the initial digital image to produce the first image comprises:

applying a directional morphological filter;

applying a sequence of multilevel morphological gradient operations; and

summing the sequence.

10. The method of claim 9, wherein applying the directional morphological filter comprises:

applying a balanced sequence of opposing morphological operations in a plurality of independent directions.

11. The method of claim 10, wherein the balance sequence of morphological operations in reverse is identical to the non-reversed sequence.

12. The method of claim 11, wherein the plurality of independent directions comprise:

a vertical direction, a horizontal direction and two opposing diagonal directions.

13. The method of claim 11, wherein the balanced sequence comprises:

a dilation operation, an erosion operation, an erosion operation, and a dilation operation.

14. The method of claim 9, wherein the sequence of multilevel operations is gradually increasing.

15. The method of claim 14, wherein the sequence of operations is increasing by two pixels in matrix width and height.

16. The method of claim 15, wherein the sequence of operations begins with an operation having a 3×

3 matrix operator.

17. An article comprising:

a storage medium;

said storage medium having stored thereon instructions to segment an initial digital image, said instructions, when executed by a computing platform, resulting in;

processing the initial digital image to produce a first digital image with defined edges corresponding to the initial digital image and a second digital image with at least two dominant contiguous regions corresponding to the initial digital image;

identifying distinct, non-overlapping regions of the first digital image formed by the defined edges;

combining the distinct, non-overlapping regions based, at least in part, on a correspondence with the at least two dominant contiguous regions of the second digital image; and

segmenting the initial digital image based, at least in part, on the remaining regions after combining the distinct, non-overlapping regions based, at least in part, on the correspondence with the at least two dominant contiguous regions.

18. The article of claim 17, wherein said instructions, when executed, resulting in:

segmenting additional digital images after said initial digital image by determining the pixel locations in a background segmentation that have changed a measurable amount in pixel signal value and processing those pixel locations as in a foreground segmentation.

19. The article of claim 17, wherein said instructions, when executed, resulting in:

segmenting additional digital images after said initial digital image by determining the pixel locations in a foreground segmentation that have not changed an amount in pixel signal value that is specifically measurable and processing those pixel locations as in a background segmentation.