Image defect correction in transform space

• US 6,393,160 B1
• Filed: 03/09/1999
• Issued: 05/21/2002
• Est. Priority Date: 03/13/1998
• Status: Expired due to Term

• Current Assignee: Kodak Alaris Inc (KPP (no. 2) Trustees Ltd.)
• Original Assignee: Applied Science Fiction, Inc.
• Inventors: Edgar, Albert D.
• Primary Examiner(s): ROGERS, SCOTT A
• Application Number:
  US09/264,773
• Time in Patent Office: 1,169 Days
• Field of Search: 382/275, 382/260, 382/263, 382/254, 358/463, 250/330, 250/559.04, 250/559.05, 250/559.06, 250/559.07, 250/559.08
• US Class Current: 382/275
• CPC Class Codes: G06T 5/50 using two or more images, e.g. averaging or subtraction
• CPC Class Codes: H04N 1/38 Circuits or arrangements for blanking or otherwise eliminating unwanted parts of pictures H04N1/387 takes precedence
• CPC Class Codes: H04N 1/4097 Removing errors due external factors, e.g. dust, scratches
• CPC Class Codes: H04N 23/81 for suppressing or minimising disturbance in the image signal generation
• CPC Class Codes: H04N 5/253 Picture signal generating by scanning motion picture films or slide opaques, e.g. for telecine scanning details therefor H04N3/36 ; standard conversion therefor H04N7/0112
• CPC Class Codes: H04N 9/11 Scanning of colour motion picture films, e.g. for telecine

Abstract

Surface defects in a reflection scan of a print made with visible light are corrected by using a scan of the print made with infrared light. In accordance with the present invention, surface defects in a reflection scan of an image consisting of pixels made with visible light are corrected by using a scan of the image consisting of pixels made with infrared light. This correction of surface defects is preformed by first establishing for each pixel an upper and lower bound for defect intensity based on the infrared record. The corresponding visible pixel is then corrected by subtracting the combination of upper and lower bound resulting in a corrected pixel.

Claims

1. A method for removing the effects of defects from an image comprising:

receiving a defective first image including a plurality of pixels, each having an intensity value;

receiving a defect image of defects in the first image including a plurality of pixels, each having an intensity value and a correspondence to the first image pixels;

selecting an element of the first image comprising a select pixel of the first image and a corresponding pixel of the defect image;

determining an upper bound for the element as a function of the defect image;

determining a lower bound for the element as a function of the defect image; and

correcting the first image as a function of the upper and lower bound.

2. The method as recited in claim 1, further including the steps of filtering the first image and filtering the second image.

3. The method as recited in claim 2, wherein the steps of filtering includes passing all high frequencies.

4. The method as recited in claim 2, wherein filtering is further refined to distinguish spatial frequencies.

5. The method as recited in claim 4, wherein a first region is defined to include a plurality of pixels of the first image and a corresponding plurality of the pixels of the defect image, and further comprising the steps of

applying a transform to the plurality of pixels from the first image to generate a plurality of elements from the transform distinguishing spatial frequency and angle; and

applying the transform to the plurality of pixels from the defect image to generate a plurality of defect elements from the transform distinguishing spatial frequency and angle.

6. The method as recited in claim 5, wherein the transform is a DCT (discrete cosine transform).

7. The method as recited in claim 5, wherein the transform is a Fourier transform.

8. The method as recited in claim 5, wherein a second region is defined that partially overlaps the first region.

9. The method as recited in claim 1, wherein the upper bound for the element is determined by multiplying the corresponding defect element by a constant.

10. The method as recited in claim 9, wherein the constant is greater than 1.0.

11. The method as recited in claim 9, wherein the lower bound for the elements is determined by multiplying the corresponding defect element by a second constant.

12. The method as recited in claim 11, wherein the second constant is between 0 and 1.0.

13. The method as recited in claim 5, wherein the upper bound for the corresponding defect element is determined by multiplying the magnitude of the defect element by a function of the spatial frequency of the corresponding element.

14. The method as recited in claim 13, wherein the function is near unity for low frequencies and rises with increasing frequency.

15. The method as recited in claim 13, wherein the lower bound for the corresponding defect element is determined by multiplying the magnitude of the defect element by a second function of the spatial frequency of the corresponding element.

16. The method as recited in claim 15, wherein the second function is near unity at low frequencies and decreases at higher frequencies.

17. The method as recited in claim 1, wherein the upper bound is a function of the select pixel of the defect image and another pixel adjacent to the select pixel.

18. The method as recited in claim 17, wherein the function includes finding the maximum of the select pixel and the adjacent pixel.

19. The method as recited in claim 1, wherein the lower bound is a function of the select pixel of the defect image and another pixel adjacent to the select pixel.

20. The method as recited in claim 19, wherein the function includes finding the minimum of the select pixel and the adjacent pixel.

21. The method as recited in claim 5, wherein a second region is defined to include a second plurality of pixels from the defect image that is offset from the first region by substantially the dimensions of an adjacent pixel, applying the transform to the second plurality of pixels to generate a second plurality of defect elements from the transform, and wherein the function determining the upper bound for the element is further characterized as comprising the step of finding the maximum of the corresponding defect element from the transform and corresponding second defect element from the second transform.

22. The method as recited in claim 21, wherein the spatial direction of offset of the second region is selected such that the correlation of the second plurality of defect elements to the plurality of elements generated from the first image is maximized.

23. The method as recited in claim 1, wherein the bounding function is further defined to comprise the steps of

selecting a ratio of the upper and lower bounds, adding the upper and lower bounds in proportion to the ratio to produce a blended bound;

subtracting the blended bound from the corresponding element of the first image to produce a corrected element; and

adjusting the ratio such that the magnitude of the element of the corrected image is minimized.

24. The method as recited in claim 23, wherein adjusting the ratio is performed by the steps of

subtracting the upper bound from the corresponding element of the first image to produce a first candidate correction;

subtracting the lower bound from the corresponding element of the second image to produce a second candidate correction;

producing a corrected element of substantially 0 if the first and second candidate corrections are of opposite sign, and a corrected element corresponding to the one of the first and second candidate corrections that is smallest in magnitude if the first and second candidate corrections are of the same sign.