Projection based method for scratch and wire removal from digital images

• US 5,974,194 A
• Filed: 10/21/1996
• Issued: 10/26/1999
• Est. Priority Date: 10/30/1995
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Totsuka, Takashi, Hirani, Anil
• Primary Examiner(s): LEE, TOMMY D
• Application Number:
  US08/733,953
• Time in Patent Office: 1,100 Days
• Field of Search: 382/254, 382/260, 382/262, 382/276, 382/280, 382/282, 382/283, 358/448, 358/463
• US Class Current: 382/262
• CPC Class Codes: G06T 2207/20056 Discrete and fast Fourier transform, [DFT, FFT]
• CPC Class Codes: G06T 5/10 using non-spatial domain filtering
• CPC Class Codes: G06T 5/77 Retouching; Inpainting; Scratch removal
• CPC Class Codes: G06V 10/30 Noise filtering

Abstract

The present invention is directed to a method of scratch removal from digitized images wherein a "scratched" portion of a digitized image which requires repair or noise removal, is identified and a binary mask is generated which distinguishes the defined "scratched" portion from the other portions of the digitized image. A repair window is defined which delineates the portion of the digitized image desired to be repaired of scratches or other noise. A sample window having the same dimensional attributes as the repair window, and preferably free of any scratch or other noise, is then defined. The sample window is preferably selected so as to be as close in resemblance to the repair window as possible. Image data of the repair window and the sample window are converted into the frequency domain and then processed to yield appropriate pixel data for transfer into the repair window so as to remove the scratch noise in such a way that alignment of edges, as well as textures and other details are maintained in the resulting digitized image without degrading existing non-scratched image areas.

Claims

1. A method of removing scratch noise from a digitized image, comprising the steps of:

inputting digitized image data;

displaying said digitized image data on a display device;

identifying a scratch noise area in said digitized image data displayed on said display device;

generating binary mask data which distinguishes pixels within the area of said identified scratch noise from pixels within a remainder of said displayed image;

storing said binary mask data;

defining a repair window area on said displayed image, wherein

said repair window area contains identified scratch noise areas;

storing data representing values of pixels within said repair window area;

defining a sample window area within said displayed image, wherein

said sample window is chosen so as to resemble the features and values of said repair window area;

storing data representing values of pixels within said sample window;

calculating a fast Fourier transform of said data representing values of pixels within said repair window;

calculating a fast Fourier transform of said data representing values of pixels within said sample window;

generating new image data in accordance with said fast Fourier transformed pixel value data of said repair window and said fast Fourier transformed pixel value data of said sample window;

calculating an inverse fast Fourier transform of said new image data;

conforming the values of said new image data to predefined limits;

generating new repair window data in accordance with said new image data, said data representing values of pixels within said repair window area and said binary mask data; and

replacing said data representing values of pixels within said repair window area with said new repair window data.

2. An apparatus for removal of scratch noise from a digitized image, comprising:

an input for receiving digitized image data;

a display device for displaying digitized image data;

scratch identifying means for identifying scratch noise within a digitized image displayed on said display device;

a processor for generating binary mask data which distinguishes identified scratch noise within a digitized image from other areas of said digitized image;

means for identifying repair window and sample window areas;

a first storage memory for storing data representing said identified repair window area;

a second storage memory for storing data representing said identified sample window area;

a third storage memory for storing said binary mask data, wherein

said processor further comprises;

means for calculating fast Fourier transform data for said repair window data and said sample window data;

means for generating new image data in accordance with the fast Fourier transform data of said sample and repair windows;

means for calculating inverse fast Fourier transform data for said new image data;

means for conforming values of said new image data to predefined limits; and

means for generating new repair window data in accordance with said new image data, repair window data and said binary mask data.

3. A method of removing scratch noise from a digitized image, comprising the steps of:

identifying a scratch noise area in said digitized image data;

defining repair window data which contains the identified scratch noise area;

defining sample window data which is chosen so as to resemble the features and values of said repair window data;

transforming said repair window data and said sample window data to data of a frequency domain;

generating new image data based upon said repair window transformed data and said sample window transformed data;

inverse-transforming said new image data;

generating new repair window data based upon said new image data; and

replacing said repair window data with said new repair window data.

4. The method of claim 3, wherein said step of transforming uses a Fourier transform.

5. The method of claim 3, further comprising the step of:

conforming values of said new image data to predefined limits.

6. The method of claim 3, wherein said new image data is generated using a DC magnitude value of a repair window frequency spectrum.

7. The method of claim 6, wherein said new image data is generated using a smaller of a magnitude of said repair window frequency spectrum and a magnitude of the sample window frequency spectrum.

8. The method of claim 3, further comprising the steps of:

generating mask data which distinguishes pixels within the area of said identified scratch noise from pixels within a remainder of said image, wherein

said new repair window data is generated based upon said new image data, repair window data and said mask data.

9. The method of claim 8, wherein said mask data is binary data.

10. An apparatus for removing scratch noise from a digitized image, comprising:

means for identifying a scratch noise area in said digitized image data;

means for defining repair window data which contains the identified scratch noise area;

means for defining sample window data which is chosen so as to resemble the features and values of said repair window data;

means for transforming said repair window data and said sample window data to data of a frequency domain;

means for generating new image data based upon said repair window transformed data and said sample window transformed data;

means for inverse-transforming said new image data;

means for generating new repair window data based upon said new image data; and

means for replacing said repair window data with said new repair window data.

11. The apparatus of claim 10, wherein means for transforming uses a Fourier transform.

12. The apparatus of claim 10, further comprising:

means for conforming values of said new image data to predefined limits.

13. The apparatus of claim 10, wherein said new image data is generated using a DC magnitude value of a repair window frequency spectrum.

14. The apparatus of claim 13, wherein said new image data is generated using a smaller of a magnitude of said repair window frequency spectrum and a magnitude of the sample window frequency spectrum.

15. The apparatus of claim 10 further comprising:

means for generating mask data which distinguishes pixels within the area of said identified scratch noise from pixels within a remainder of said image, wherein

said new repair window data is generated based upon said new image data, repair window data and said mask data.

16. The apparatus of claim 15, wherein said mask data is binary data.

17. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform method steps for removing scratch noise from a digitized image, the method steps comprising:

identifying a scratch noise area in said digitized image data;

defining repair window data which contains the identified scratch noise area;

defining sample window data which is chosen so as to resemble the features and values of said repair window data;

transforming said repair window data and said sample window data to data of a frequency domain;

generating new image data based upon said repair window transformed data and said sample window transformed data;

inverse-transforming said new image data;

generating new repair window data based upon said new image data; and

replacing said repair window data with said new repair window data.

18. The program storage device of claim 17, wherein said step of transforming uses a Fourier transform.

19. The program storage device of claim 17, wherein the method further comprises the step of:

conforming values of said new image data to predefined limits.

20. The program storage device of claim 17, wherein said new image data is generated using a DC magnitude value of a repair window frequency spectrum.

21. The program storage device of claim 20, wherein said new image data is generated using a smaller of a magnitude of said repair window frequency spectrum and a magnitude of the sample window frequency spectrum.

22. The program storage device of claim 17, wherein the method further comprises the steps of:

generating mask data which distinguishes pixels within the area of said identified scratch noise from pixels within a remainder of said image, wherein

said new repair window data is generated based upon said new image data, repair window data and said mask data.

23. The program storage device of claim 22, wherein said mask data is binary data.

24. A recording medium having digital data recorded thereon, the recording medium being prepared by the steps of:

identifying a scratch noise area in said digitized image data;

defining repair window data which contains the identified scratch noise area;

defining sample window data which is chosen so as to resemble the features and values of said repair window data;

transforming said repair window data and said sample window data to data of a frequency domain;

generating new image data based upon said repair window transformed data and said sample window transformed data;

inverse-transforming said new image data;

generating new repair window data based upon said new image data;

replacing said repair window data with said new repair window data; and

recording said new repair window data on the recording medium.

25. The recording medium of claim 24, wherein said step of transforming uses a Fourier transform.

26. The recording medium of claim 24, further comprising the step of:

conforming values of said new image data to predefined limits.

27. The recording medium of claim 24, wherein said new image data is generated using a DC magnitude value of a repair window frequency spectrum.

28. The recording medium of claim 27, wherein said new image data is generated using a smaller of a magnitude of said repair window frequency spectrum and a magnitude of the sample window frequency spectrum.

29. The recording medium of claim 24, further comprising the steps of:

generating mask data which distinguishes pixels within the area of said identified scratch noise from pixels within a remainder of said image, wherein

said new repair window data is generated based upon said new image data, repair window data and said mask data.

30. The recording medium of claim 29, wherein said mask data is binary data.

31. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform method steps for removing scratch noise from a digitized image, the method steps comprising:

inputting digitized image data;

displaying said digitized image data on a display device;

identifying a scratch noise area in said digitized image data displayed on said display device;

generating binary mask data which distinguishes pixels within the area of said identified scratch noise from pixels within a remainder of said displayed image;

storing said binary mask data;

defining a repair window area on said displayed image, wherein

said repair window area contains identified scratch noise areas;

storing data representing values of pixels within said repair window area;

defining a sample window area within said displayed image, wherein

said sample window is chosen so as to resemble the features and values of said repair window area;

storing data representing values of pixels within said sample window;

calculating a fast Fourier transform of said data representing values of pixels within said repair window;

calculating a fast Fourier transform of said data representing values of pixels within said sample window;

generating new image data in accordance with said fast Fourier transformed pixel value data of said repair window and said fast Fourier transformed pixel value data of said sample window;

calculating an inverse fast Fourier transform of said new image data;

conforming the values of said new image data to predefined limits;

generating new repair window data in accordance with said new image data, said repair window data and said binary mask data; and

replacing said repair window data with said new repair window data.

32. A recording medium having digital data recorded thereon, the recording medium being prepared by the steps of:

inputting digitized image data;

displaying said digitized image data on a display device;

identifying a scratch noise area in said digitized image data displayed on said display device;

generating binary mask data which distinguishes pixels within the area of said identified scratch noise from pixels within a remainder of said displayed image;

storing said binary mask data;

defining a repair window area on said displayed image, wherein

said repair window area contains identified scratch noise areas;

storing data representing values of pixels within said repair window area;

defining a sample window area within said displayed image, wherein

said sample window is chosen so as to resemble the features and values of said repair window area;

storing data representing values of pixels within said sample window;

calculating a fast Fourier transform of said data representing values of pixels within said repair window;

calculating a fast Fourier transform of said data representing values of pixels within said sample window;

generating new image data in accordance with said fast Fourier transformed pixel value data of said repair window and said fast Fourier transformed pixel value data of said sample window;

calculating an inverse fast Fourier transform of said new image data;

conforming the values of said new image data to predefined limits;

generating new repair window data in accordance with said new image data, said repair window data and said binary mask data;

replacing said repair window data with said new repair window data; and

recording the new repair window data on the recording medium.