Method and apparatus for correcting pixel values in a digital image

• US 6,104,839 A
• Filed: 07/10/1996
• Issued: 08/15/2000
• Est. Priority Date: 10/16/1995
• Status: Expired due to Term

• Current Assignee: Eastman Kodak Company
• Original Assignee: Eastman Kodak Company
• Inventors: Gray, Robert Terry, Cok, David Roy, Matraszek, Tomasz Andrzej
• Primary Examiner(s): Lee, Thomas D.
• Application Number:
  US08/675,320
• Time in Patent Office: 1,497 Days
• Field of Search: 382/254, 382/275, 382/190, 382/195, 382/205, 358/447, 358/448, 358/463
• US Class Current: 382/254
• CPC Class Codes: E05F 15/63 operated by swinging arms
• CPC Class Codes: E05F 15/77 using wireless control
• CPC Class Codes: E05Y 2201/216 Clutches
• CPC Class Codes: E05Y 2201/236 using force or torque
• CPC Class Codes: E05Y 2201/246 by auxiliary motors, magnets, springs or weights
• CPC Class Codes: E05Y 2201/434 Electromotors; Details thereof
• CPC Class Codes: E05Y 2201/462 Electromagnets
• CPC Class Codes: E05Y 2400/612 Batteries
• CPC Class Codes: E05Y 2600/46 in or on the wing
• CPC Class Codes: E05Y 2800/00 Details, accessories and auxiliary operations not otherwise provided for
• CPC Class Codes: E05Y 2900/132 Doors
• CPC Class Codes: G06T 5/20 using local operators

Abstract

A method and apparatus for editing continuous tone digital images representing pixels in which regions containing defects, called "defect regions", are unnoticeable or less visually discernible. Corrected pixel data for pixels in the defect region, called defect pixels, in the digital image are provided while data representing non-defect pixels outside of the defect region are not altered. These digital images may have one color channel (gray side) or multiple color channels. The apparatus and method includes a computer system which operates according to a set of instructions.

Claims

1. A method for correcting one or more defect pixels in a source image with non-defect pixels and the defect pixels, each of the defect pixels and the non-defect pixels being represented by at least one defect pixel signal and non-defect pixel signal, respectively, said method comprising the steps of:

selecting one of said defect pixels;

allocating a plurality of line segments through said selected defect pixel, wherein each said line segment is composed of the defect pixels and the non-defect pixels about said selected defect pixel;

determining at least one representative model of the non-defect pixel signals along each said line segment;

determining the deviation of the non-defect pixel signals along each said line segment from the model corresponding to the line segment;

selecting the line segment having the lowest deviation; and

determining corrected pixel signals for said selected defect pixel based on at least two of the non-defect pixel signals in said selected line segment.

2. A method according to claim 1 wherein one or more of said defect pixels are located within a defect region of said source image, and said step of

selecting one of said defect pixels includes selecting one of said defect pixels in said defect region.

3. A method according to claim 1 wherein:

said defect pixel signals represent channel signals of their corresponding defect pixels;

said non-defect pixel signals represent channel signals of their corresponding non-defect pixels;

said model determining step further comprises determining a representative model based upon corresponding channel signals of the non-defect pixels along each said line segment; and

said deviation determining step further comprises determining a deviation of corresponding channel signals of the non-defect pixels along each said line segment from the model of said channel signals of said non-defect pixels along the line segment.

4. A method according to claim 3 wherein:

said deviation determining step further comprises the step of determining a total for each said line segment based on the deviation of the channel signals of the non-defect pixels;

said line segment selecting step further comprises the step of selecting the line segment having the lowest total; and

said pixel signals determining step further comprises determining corrected channel signals for said selected defect pixel based on the channel signals of at least two of said non-defect pixels in said selected line segment.

5. A method according to claim 1 further comprising the step of scanning successively the defect pixels and the non-defect pixels in said source image, and said defect pixel selecting step further comprises the step of selecting each said defect pixel in said source image during said successive scanning of said defect pixels in said source image.

6. A method according to claim 5 further comprising the step of storing in a corrected image the pixel signals of said non-defect pixels in said source image, and the corrected pixel signals for each said selected defect pixel.

7. A method according to claim 6 further comprising the step of outputting the pixel signals of said corrected image.

8. A method according to claim 1 wherein said allocating step is carried out to provide the line segments at equal radial angles about said selected defect pixel, each said line segment being composed of the defect pixels and the non-defect pixels in said source image near the respective radial angle of the line segment.

9. A method according to claim 1 wherein said allocating step further comprises the step of allocating each line segment through said selected defect pixel such that the defect pixels and the non-defect pixels in the source image composing said line segment are from opposite sides of said selected defect pixel.

10. A method according to claim 9 wherein said model determining step and said deviation determining step is performed only on line segments composed substantially of non-defect pixels on each said opposite side of said selected defect pixel.

11. A method according to claim 9 wherein said allocating step further comprises composing each said line segment of a predefined number of the non-defect pixels on each said opposite side of said selected defect pixel.

12. A method according to claim 9 wherein said allocating step further comprises the step of composing each said line segment of a variable number of the defect pixels and the non-defect pixels by extending the line segment along a vector from each said opposite side of said selected defect pixel, wherein said variable number for each said line segment is based upon statistical consistency of the non-defect pixel signals along the vector.

13. A method according to claim 12 wherein said composing step further comprises the step of determining said variable number of pixels for each said opposite side by

placing said non-defect pixels for said side into groups of adjacent non-defect pixels,

selecting a first one of said groups, and a second one of said groups, wherein said second group comprises the non-defect pixels in said first group and the group next furthest from said selected defect pixel which is adjacent to said first group,

calculating a first sum mean square error of the signals of the non-defect pixels in said first group and a second sum mean square error of the signals of the non-defect pixels in said second group,

calculating a ratio of said second sum mean square error to the greater of said first sum mean square error or a second threshold, and

fixing said variable number to the number of the defect pixels and non-defect pixels from said selected pixel to said non-defect pixel in said first group furthest from said selected defect pixel when one of

said second sum mean square error is greater than said first threshold, and

said ratio is greater than a third threshold, wherein said first threshold is greater than said second threshold, and said third threshold is greater than one.

14. A method according to claim 13 wherein said step of determining said variable number of pixels for each said opposite side further comprises the steps of:

fixing said variable number to the number of the defect pixels and the non-defect pixels to the last non-defect pixel in said first group within said source image when one of said first and said second groups extends outside said source image; and

fixing said variable number to the number of the defect pixels and the non-defect pixels to the last non-defect pixel in said first group when said non-defect pixels in said first group exceed a maximum threshold.

15. A method according to claim 9 wherein said corrected pixel signals determining step further comprises interpolating pixel signals for said selected defect pixel based on at least one of the non-defect pixel signals in said selected line segment from each said opposite side of said selected defect pixel.

16. A method according to claim 1 wherein said allocating step further comprises composing said line segments of the same total number of the defect pixels and non-defect pixels.

17. A method according to claim 1 wherein said deviation determining step further comprises the steps for each line segment of:

determining estimated signals for each non-defect pixel in the line segment based upon one of the models of the non-defect pixel signals of the line segment;

determining an error signal based upon the difference between the non-defect pixel signals and the estimated signals of each non-defect pixel in the line segment;

determining a square of each error signal; and

determining a mean of the squared error signals to provide said deviation.

18. A method according to claim 1 wherein said step of determining the deviation of said pixel signals further comprises the step for each line segment of:

determining a standard deviation of the pixel signals of the non-defect pixels in the line segment, wherein said deviation of the line segment is said standard deviation.

19. An apparatus for correcting values of defect pixels in a source image having a programmed computer with memory for storing said source image, said stored source image comprising a plurality of pixels having signals representing values of one or more color channels, and a defect region comprising the pixels within said plurality having corrupted values, wherein said pixels in said defect region are defect pixels, and said pixels outside said defect region are non-defect pixels, said apparatus comprising:

means for selecting one of said defect pixels in said defect region;

means for allocating a plurality of line segments through said selected defect pixel, wherein each said line segment is composed of pixels in said source image about said selected defect pixel;

means for determining for each channel of each said line segment a representative model of the non-defect pixels values;

means for determining for each channel of each said line segment the deviation of the non-defect pixels values in said line segment from their representative model;

means for determining a total for each line segment based on said deviation for each channel of said line segment;

means for selecting the line segment having the lowest said total deviation from among said plurality of line segments; and

means for determining a value for each channel of said selected defect pixel based on the values of at least two of said non-defect pixels in said selected line segment.

20. An apparatus according to claim 19 further comprising means for scanning successively each pixel in said source image, wherein said selecting means selects said defect pixel during said scanning when said scanned pixel is one of said defect pixels in said source image.

21. An apparatus according to claim 20 further comprising means for storing in said memory during said scanning by said scanning means of each pixel in said source image a corrected image having signals representing the value of each channel of said non-defect pixels in said source image, and the determined value for each channel of each said selected defect pixel.

22. An apparatus according to claim 21 further comprising an output device coupled to said computer, and means for outputting from said memory to said output device said signals representing the values of the pixels of said stored corrected image in said memory.

23. An apparatus for correcting one or more defect pixels in a source image with non-defect pixels and the defect pixels, each of the defect pixels and the non-defect pixels being represented by at least one defect pixel signal and non-defect pixel signal, respectively, said apparatus comprising:

means for selecting one of said defect pixels;

means for allocating a plurality of line segments through said selected defect pixel, wherein each said line segment is composed of the defect pixels and the non-defect pixels about said selected defect pixel;

means for determining at least one representative model of the non-defect pixel signals along each said line segment;

means for determining the deviation of the non-defect pixel signals along each said line segment from the model corresponding to the line segment;

means for selecting the line segment having the lowest deviation; and

means for determining corrected pixel signals for said selected defect pixel based on at least two of the non-defect pixel signals in said selected line segment.

24. An apparatus for correcting values of defect pixels in a source image having a programmed computer with memory for storing said source image, said stored source image comprising a plurality of pixels having signals representing values of one or more color channels, and a defect region comprising the pixels within said plurality having corrupted values, wherein said pixels in said defect region are defect pixels, and said pixels outside said defect region are non-defect pixels, said apparatus further comprising said programmed computer having a processor wherein:

said processor selects one of said defect pixels in said defect region;

allocates a plurality of line segments through said selected defect pixel wherein each said line segment is composed of pixels in said source image about said selected defect pixel, determines for each channel of each said line segment a representative model of the non-defect pixels values, determines for each channel of each said line segment the deviation of the non-defect pixels values in said line segment from their representative model, determines a total for each line segment based on said deviation for each channel of each line segment, selects the line segment having the lowest said total from among said plurality of line segments, and determines a value for each channel of said selected defect pixel based on the values of at least two of said non-defect pixels in said selected line segment.

25. An apparatus according to claim 24 further comprising:

An output device coupled to said processor; and

said processor also scans successively each pixel in said source image, wherein said selection of said defect pixel occurs during one or more of said successive scans, stores in said memory a corrected image having signals representing the value of each channel of said non-defect pixels in said source image and the determined value for each channel of each said selected defect pixel, and outputs from said memory to said output device the signals representing values of the pixels of said corrected image.