Image processing system, mask fabrication method, and program

• US 8,311,332 B2
• Filed: 08/31/2006
• Issued: 11/13/2012
• Est. Priority Date: 09/06/2005
• Status: Active Grant

• Current Assignee: mofiria Corp.
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Abe, Hiroshi
• Primary Examiner(s): Akhavannik, Hadi
• Application Number:
  US11/718,784
• Publication Number:
  US 20090129635A1
• Time in Patent Office: 2,266 Days
• Field of Search: 382/190, 382/177, 382/100, 382/115
• US Class Current: 382/177
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06T 2207/10048 Infrared image
• CPC Class Codes: G06T 2207/30101 Blood vessel; Artery; Vein; Vascular
• CPC Class Codes: G06T 2207/30196 Human being; Person
• CPC Class Codes: G06T 7/12 Edge-based segmentation
• CPC Class Codes: G06V 10/26 Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
• CPC Class Codes: G06V 10/46 Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features colour feature extraction G06V10/56
• CPC Class Codes: G06V 40/12 Fingerprints or palmprints
• CPC Class Codes: G06V 40/14 Vascular patterns

Abstract

An image processing system and a mask preparation method able to prepare a mask by simple processing and a program executed in such an image processing system are provided. To extract the edges of the image, strings of pixels corresponding to the contours of an object are extracted from the edge extracted image, and border lines for the masking are acquired based on an approximation line thereof.

Claims

1. An image processing system for preparing a mask for cutting out an image of a predetermined object from an image, said image processing system comprising:

an edge extraction unit configured to extract the edges of said image;

a pixel extraction unit configured to extract a string of first pixels included in a first region contiguous with a first reference line set on a plane of said image so as to pass through an inside of the image of said object, located on a plurality of parallel second reference lines set on the plane of said image, and nearest to said first reference line on the respective plurality of second reference lines and a string of second pixels included in a second region contiguous with said first reference line on an opposite side of said first region, located on said plurality of second reference lines, and nearest said first reference line on the respective plurality of second reference lines from among edges extracted at said edge extraction unit;

a first intermediate point extraction unit configured to extract a string of intermediate points located between said first pixels and said second pixels on the respective plurality of second reference lines and in which a distance from the first pixels and the distance from the second pixels have a predetermined ratio;

a division unit configured to divide the string of intermediate points extracted at said first intermediate point extraction unit into at least one section in which adjacent intermediate points are within a predetermined distance of each other;

an approximation line acquisition unit configured to acquire an approximation line of the string of intermediate points extracted at said first intermediate point extraction unit based on the intermediate points forming the longest section divided at said division unit;

a second intermediate point extraction unit configured to extract the string of intermediate points having a difference from the approximation line acquired at said approximation line acquisition unit within a predetermined range from among intermediate points extracted at said first intermediate point extraction unit; and

a first border line acquisition unit configured to acquire a first border line of said mask based on an approximation line of the string of first pixels located on the same second reference lines as the string of intermediate points extracted at said second intermediate point extraction unit and acquire a second border line of said mask based on an approximation line of the string of second pixels located on the second reference lines.

2. An image processing system as set forth in claim 1, wherein:

said first border line acquisition unit acquires a line obtained by shifting the approximation line of the string of said first pixels toward said first reference line by exactly a predetermined distance as said first border line and acquires a line obtained by shifting the approximation line of the string of said second pixels toward said first reference line by exactly a predetermined distance as said second border line.

3. An image processing system as set forth in claim 1, comprising a third border line acquisition unit for sequentially moving a region surrounded by two sides parallel to said first reference line and two sides parallel to said second reference lines and including a portion of said first border line and a portion of said second border line from a predetermined position on said image in one direction parallel with respect to said first reference line, counting the number of pixels of edges extracted at said edge extraction unit included in the region at each moved position, and acquiring a line parallel to said second reference line which passes through the region at the position where the count reaches the predetermined threshold value as a third border line of said mask.

4. An image processing system as set forth in claim 1, wherein:

where said pixel extraction unit cannot extract at least one of said first pixels and said second pixels in a portion of said plurality of second reference lines, said first intermediate point extraction unit interpolates intermediate points of the portion of the second reference lines based on intermediate points extracted on other second reference lines adjacent to the second reference lines of that portion.

5. An image processing system as set forth in claim 1, wherein:

said first border line acquisition unit acquires said first border line and said second border line based on the intermediate points extracted at said second intermediate point extraction unit and the interval of first pixels and second pixels located on the same second reference lines sandwiching the extracted intermediate points therebetween.

6. An image processing system as set forth in claim 1, further comprising:

an edge enhancing unit configured to enhance the edges of said image extracted at said edge extraction unit,

a region specifying unit configured to specify a region further inside of the object cut out by said mask,

an evaluation value acquisition unit configured to acquire an evaluation value concerning an intensity of edges and/or an amount of edges included in the region specified at said region specifying unit based on values of pixels included in the region at said image enhanced in edges at said edge enhancing unit, and

a judgment unit configured to judge whether or not the image of said object is cut out by said mask based on the evaluation value acquired at said evaluation value acquisition unit.

7. An image processing system for preparing a mask for cutting out an image of a predetermined object from an image, said image processing system comprising:

an edge extraction unit configured to extract edges of said image;

a first pixel extraction unit configured to extract a string of pixels included in one of two regions contiguous with a first reference line set on a plane of said image so as to pass through the inside of the image of said object, located on a plurality of parallel second reference lines set on the plane of said image, and nearest to said first reference line on the respective plurality of second reference lines from among edges extracted at said edge extraction unit;

a division unit configured to divide the string of pixels extracted at said first pixel extraction unit into at least one section in which adjacent pixels are within a predetermined distance of each other;

an approximation line acquisition unit configured to acquire an approximation line of the string of pixels extracted at said first pixel extraction unit based on pixels forming the longest section divided at said division unit;

a second pixel extraction unit configured to extract a string of pixels having a difference from the approximation line acquired at said approximation line acquisition unit within a predetermined range among pixels extracted at said first pixel extraction unit; and

a first border line acquisition unit configured to acquire a first border line of said mask based on the approximation line of the string of pixels extracted at said second pixel extraction unit.

8. An image processing system as set forth in claim 7, further comprising a second border line acquisition unit configured to move a strip shaped region sandwiched by two parallel sides in parallel with respect to said second reference lines, count the number of pixels of edges extracted at said edge extraction unit included in the region at each position of movement, and acquire a second border line of said mask in accordance with the line parallel to said sides included in the region at a position where the count becomes the largest, and

said acquired second border line being set as said first reference line.

9. An image processing system as set forth in claim 8, wherein:

said first border line acquisition unit acquires a line obtained by shifting the approximation line of the string of pixels extracted at said second pixel extraction unit toward said second border line by exactly a predetermined distance as said first border line, and

said second border line acquisition unit acquires a line obtained by shifting the line parallel to said sides included at said region at the position where the count of the number of pixels of said edges becomes the largest toward said first border line by exactly a predetermined distance as said second border line.

10. An image processing system as set forth in claim 8, further comprising a third border line acquisition unit configured to sequentially move a region surrounded by two sides parallel to said first reference line and two sides parallel to said second reference lines and including a portion of said first border line and a portion of said second border line from a predetermined position on said image in one direction parallel with respect to said first reference line, count the number of pixels of edges extracted at said edge extraction unit included in the region at each moved position, and acquire a line parallel to said second reference line which passes through the region at the position where the count reaches a predetermined threshold value as a third border line of said mask.

11. An image processing system as set forth in claim 8, further comprising:

an edge enhancing unit configured to enhance the edges extracted at said edge extraction unit,

a region specifying unit configured to specify a region further inside of the object cut out by said mask,

an evaluation value acquisition unit configured to acquire an evaluation value concerning an intensity of edges and/or an amount of edges included in the region specified at said region specifying unit based on values of pixels included in the region at said image enhanced in edges at said edge enhancing unit, and

a judgment unit configured to judge whether or not an image of said object is cut out by said mask based on the evaluation value acquired at said evaluation value acquisition unit.

12. A mask preparation method for preparing a mask for cutting out an image of a predetermined object from an image, said mask preparation method including:

a first step of extracting the edges of said image;

a second step of extracting a string of first pixels included in a first region contiguous with a first reference line set on a plane of said image so as to pass through an inside of the image of said object, located on a plurality of parallel second reference lines set on the plane of said image, and nearest to said first reference line on the respective plurality of second reference lines and a string of second pixels included in a second region contiguous with said first reference line on an opposite side of said first region, located on said plurality of second reference lines, and nearest to said first reference line on the respective plurality of second reference lines from among edges extracted at said first step;

a third step of extracting a string of intermediate points located between said first pixels and said second pixels on said plurality of second reference lines and in which a distance from the first pixel and the distance from the second pixel have a predetermined ratio;

a fourth step of dividing the string of intermediate points extracted at said third step into one or more sections in which adjacent intermediate points are within a predetermined distance of each other;

a fifth step of acquiring an approximation line of the string of intermediate points extracted at said third step based on the intermediate points forming the longest section divided at said fourth step;

a sixth step of extracting a string of intermediate points having a difference from the approximation line acquired at said fifth step within a predetermined range from among intermediate points extracted at said third step; and

a seventh step of acquiring a first border line of said mask based on the approximation line of the string of first pixels located on the same second reference lines as the string of intermediate points extracted at said sixth step and acquiring a second border line of said mask based on the approximation line of the string of second pixels located on the second reference lines.

13. A mask preparation method for preparing a mask for cutting out an image of a predetermined object from an image, said mask preparation method including:

a first step of extracting edges of said image;

a second step of extracting a string of first pixels included in one of two regions contiguous with the first reference line set on the plane of said image so as to pass through the inside of the image of said object, located on a plurality of parallel second reference lines set on the plane of said image, and nearest to said first reference line on the respective plurality of second reference lines from among edges extracted at said first step;

a third step of dividing the string of first pixels extracted at said second step into at least one section in which adjacent first pixels are within a predetermined distance of each other;

a fourth step of acquiring an approximation line of the string of first pixels extracted at said second step based on first pixels forming the longest section divided at said third step;

a fifth step of extracting a string of first pixels having a difference from the approximation line acquired at said fourth step within a predetermined range among first pixels extracted at said second step; and

a sixth step of acquiring a first border line of said mask based on the approximation line of the string of first pixels extracted at said fifth step.

14. A mask preparation method as set forth in claim 13, further including a seventh step of moving a strip shaped region sandwiched by two parallel sides in parallel with respect to said second reference lines, counting the number of pixels of edges extracted at said first step included in the region at each position of the movement, and acquiring a second border line of said mask in accordance with the line parallel to said sides included in the region at a position where the count becomes the largest, and

said acquired second border line being set as said first reference line.

15. A program on a non-transitory computer-readable medium for making an image processing system having a computer for preparing a mask for cutting out an image of a predetermined object from an image execute:

a first step of extracting the edges of said image;

a second step of extracting a string of first pixels included in a first region contiguous with a first reference line set on a plane of said image so as to pass through an inside of the image of said object, located on a plurality of parallel second reference lines set on the plane of said image, and nearest to said first reference line on the respective plurality of second reference lines and a string of second pixels included in a second region contiguous with said first reference line on an opposite side of said first region, located on said plurality of second reference lines, and nearest to said first reference line on the respective plurality of second reference lines from among edges extracted at said first step;

a third step of extracting a string of intermediate points located between said first pixels and said second pixels on said plurality of second reference lines and in which a distance from the first pixel and the distance from the second pixel have a predetermined ratio;

a fourth step of dividing the string of intermediate points extracted at said third step into one or more sections in which adjacent intermediate points are within a predetermined distance of each other;

a fifth step of acquiring an approximation line of the string of intermediate points extracted at said third step based on the intermediate points forming the longest section divided at said fourth step;

a sixth step of extracting a string of intermediate points having a difference from the approximation line acquired at said fifth step within a predetermined range from among intermediate points extracted at said third step; and

a seventh step of acquiring a first border line of said mask based on the approximation line of the string of first pixels located on the same second reference lines as the string of intermediate points extracted at said sixth step and acquiring a second border line of said mask based on the approximation line of the string of second pixels located on the second reference lines.

16. A program on a non-transitory computer-readable medium for making an image processing system having a computer for preparing a mask for cutting out an image of a predetermined object from an image execute:

a first step of extracting edges of said image;

a second step of extracting a string of first pixels included in one of two regions contiguous with the first reference line set on the plane of said image so as to pass through the inside of the image of said object, located on a plurality of parallel second reference lines set on the plane of said image, and nearest to said first reference line on the respective plurality of second reference lines from among edges extracted at said first step;

a third step of dividing the string of first pixels extracted at said second step into at least one section in which adjacent first pixels are within a predetermined distance of each other;

a fourth step of acquiring an approximation line of the string of first pixels extracted at said second step based on first pixels forming the longest section divided at said third step;

a fifth step of extracting a string of first pixels having a difference from the approximation line acquired at said fourth step within a predetermined range among first pixels extracted at said second step; and

a sixth step of acquiring a first border line of said mask based on the approximation line of the string of first pixels extracted at said fifth step.

17. A program as set forth in claim 16, further

making said image processing system execute a seventh step of moving a strip shaped region sandwiched by two parallel sides in parallel with respect to said second reference lines, counting the number of pixels of edges extracted at said first step included in the region at each position of the movement, and acquiring a second border line of said mask in accordance with the line parallel to said sides included in the region at a position where the count becomes the largest,

and said acquired second border line being set as said first reference line.

18. An image processing system for preparing a mask for cutting out an image of a predetermined object from an image, said image processing system comprising:

an edge extracting means for extracting edges of said image;

a pixel extracting means for extracting a string of first pixels included in a first region contiguous with a first reference line set on a plane of said image so as to pass through an inside of the image of said object, located on a plurality of parallel second reference lines set on the plane of said image, and nearest to said first reference line on the respective plurality of second reference lines and a string of second pixels included in a second region contiguous with said first reference line on an opposite side of said first region, located on said plurality of second reference lines, and nearest to said first reference line on the respective plurality of second reference lines from among edges extracted at said edge extracting means;

a first intermediate point extracting means for extracting a string of intermediate points located between said first pixels and said second pixels on said plurality of second reference lines and in which a distance from the first pixels and a distance from the second pixels have a predetermined ratio;

a dividing means for dividing the string of intermediate points extracted at said first intermediate point extracting means into at least one section in which adjacent intermediate points are within a predetermined distance of each other;

an approximation line acquiring means for acquiring an approximation line of the string of intermediate points extracted at said first intermediate point extracting means based on the intermediate points forming the longest section divided at said dividing means;

a second intermediate point extracting means for extracting a string of intermediate points having a difference from the approximation line acquired at said approximation line acquiring means within a predetermined range from among intermediate points extracted at said first intermediate point extracting means; and

a first border line acquiring means for acquiring a first border line of said mask based on the approximation line of the string of first pixels located on the same second reference lines as the string of intermediate points extracted at said second intermediate point extracting means and acquiring a second border line of said mask based on the approximation line of the string of second pixels located on the second reference lines.

19. An image processing system as set forth in claim 18, further having

a third border line acquiring means for sequentially moving a region surrounded by two sides parallel to said first reference line and two sides parallel to said second reference lines and including a portion of said first border line and a portion of said second border lines from a predetermined position on said image in one direction parallel with respect to said first reference line, counting the number of pixels of edges extracted at said edge extracting means included in the region at each moved position, and acquiring a line parallel to said second reference line which passes through the region at the position where the count reaches a predetermined threshold value as a third border line of said mask.

20. An image processing system as set forth in claim 18, further comprising:

an edge enhancing means for enhancing the edges of said image extracted at said edge extracting means,

a region specifying means for specifying a region further inside of the object cut out by said mask,

an evaluation value acquiring means for acquiring an evaluation value concerning the intensity of edges and/or the amount of edges included in the region based on the values of pixels included in the region specified at said region specifying means at said image enhanced in edges at said edge enhancing means, and

a judging means for judging whether or not the image of said object is cut out by said mask based on the evaluation value acquired at said evaluation value acquiring means.

21. An image processing system for preparing a mask for cutting out an image of a predetermined object from an image, said image processing system comprising:

an edge extracting means for extracting edges of said image;

a first pixel extracting means for extracting a string of pixels included in one of two regions contiguous with a first reference line set on a plane of said image so as to pass through the inside of the image of said object, located on a plurality of parallel second reference lines set on the plane of said image, and nearest to said first reference line on the respective plurality of second reference lines from among edges extracted at said edge extracting means;

a dividing means for dividing the string of pixels extracted at said first pixel extracting means into at least one section in which adjacent pixels are within a predetermined distance of each other;

an approximation line acquiring means for acquiring an approximation line of the string of pixels extracted at said first pixel extracting means based on pixels forming the longest section divided at said dividing means;

a second pixel extracting means for extracting a string of pixels having a difference from the approximation line acquired at said approximation line acquiring means within a predetermined range among pixels extracted at said first pixel extracting means; and

a first border line acquiring means for acquiring a first border line of said mask based on the approximation line of the string of pixels extracted at said second pixel extracting means.

22. An image processing system as set forth in claim 21, further comprising a second border line acquiring means for moving a strip shaped region sandwiched by two parallel sides in parallel with respect to said second reference lines, counting the number of pixels of edges extracted at said edge extracting means included in the region at each position of the movement, and acquiring a second border line of said mask in accordance with the line parallel to said sides included in the region at a position where the count becomes the largest, and

said acquired second border line being set as said first reference line.

23. An image processing system as set forth in claim 22, further comprising a third border line acquiring means for sequentially moving a region surrounded by two sides parallel to said first reference line and two sides parallel to said second reference lines and including a portion of said first border line and a portion of said second border line from a predetermined position on said image in one direction parallel with respect to said first reference line, counting the number of pixels of edges extracted at said edge extracting means included in the region at each moved position, and acquiring the line parallel to said second reference line which passes through the region at the position where the count reaches the predetermined threshold value as a third border line of said mask.

24. An image processing system as set forth in claim 22, further comprising:

an edge enhancing means for enhancing edges extracted at said edge extracting means,

a region specifying means for specifying a region further inside of the object cut out by said mask,

an evaluation value acquiring means for acquiring an evaluation value concerning the intensity of edges and/or the amount of edges included in the region based on the value of pixels included in the region specified at said region specifying means at said image enhanced in edges at said edge enhancing means, and

a judging means for judging whether or not the image of said object is cut out by said mask based on the evaluation value acquired at said evaluation value acquiring means.