Method and apparatus for processing digital images to suppress their noise and enhancing their sharpness

• US 6,055,340 A
• Filed: 02/27/1998
• Issued: 04/25/2000
• Est. Priority Date: 02/28/1997
• Status: Expired due to Term

• Current Assignee: Fujifilm Corporation (Fujifilm Holdings Corporation)
• Original Assignee: Fuji Photo Film Co Limited (Fujifilm Holdings Corporation)
• Inventors: Nagao, Kimitoshi
• Primary Examiner(s): Rogers, Scott
• Assistant Examiner(s): Thornton, Lartherrio E.
• Application Number:
  US09/032,112
• Time in Patent Office: 788 Days
• Field of Search: 382/260, 382/261, 382/263, 382/264, 382/266, 382/269
• US Class Current: 382/261
• CPC Class Codes: G06T 2207/20204 Removing film grain; Adding simulated film grain
• CPC Class Codes: G06T 5/75 Unsharp masking

Abstract

A method of processing digital images to suppress their noise and enhance their sharpness. The method of suppressing noise and enhancing their sharpness includes steps of: (a) performing a sharpness enhancing process on original image data to generate sharpness enhance image data also performing a smoothing process on original image data to generate smoothed image data; (b) subtracting the smoothed image data from sharpness enhanced image data to extract a component consisting of both edges and noise; (c) performing nonlinear transformation on edge/noise containing component to separate a noise component and performing a subdividing process on the resulting noise component to produce a subdivided noise component; (d) separately detecting an edge component from original image data and determining weighting data for an edge region and a noise region from the detected edge component; (e) weighting edge/noise containing component with the weighting data for edge region to produce an edge enhancing component; and (f) subtracting subdivided noise component multiplied by a factor from sharpness enhanced image data while adding edge enhance component multiplied by another factor to sharpness enhanced image data so as to generate processed image data. The effect of the method is to process digital images to suppress their noise and enhance their sharpness.

Claims

1. A method of processing digital images to suppress their noise and enhance their sharpness, comprising the steps of:

performing a sharpness enhancing process on original image data to generate sharpness enhanced image data;

performing a smoothing process on said original image data to generate smoothed image data;

subtracting the smoothed image data from said sharpness enhanced image data to extract a component consisting of both edges and noise;

performing nonlinear transformation on said edge/noise containing component to separate a noise component and performing a subdividing process on the resulting noise component to produce a subdivided noise component;

separately detecting an edge component from said original image data and determining weighting data for an edge region and a noise region from the detected edge component;

weighting said edge/noise containing component with the weighting data for said edge region to produce an edge enhancing component; and

subtracting said subdivided noise component multiplied by a factor from said sharpness enhanced image data while adding said edge enhanced component multiplied by another factor to said sharpness enhanced image data so as to generate processed image data.

2. The method according to claim 1, wherein the nonlinear transformation for separating said noise component is performed after said edge/noise containing component is weighted with the weighting data for said noise region.

3. The method according to claim 2, wherein said edge component for determining said weighting data for the edge and noise regions is directly detected from said original image data by a local variance method.

4. The method according to claim 1, wherein said edge component for determining said weighting data for the edge and noise regions is detected when said edge/noise containing component is directly processed by nonlinear transformation to separate said noise component.

5. The method according to claim 4, wherein said edge component separated from said edge/noise containing component is smoothed with a median filter for removing graininess before the weighting data for said edge and noise regions are determined.

6. The method according to claim 1, wherein said subdivided noise component multiplied by the factor is subtracted from said sharpness enhanced image data after it is weighted by the weighting data for said noise region.

7. The method according to claim 1, wherein the nonlinear transformation for separating said noise component is performed by a function which decreases exponentially in a continuous manner as the density variation of noise either increases or decreases from its threshold.

8. The method according to claim 1, wherein the nonlinear transformation for separating said noise component is performed by a function which decreases exponentially in a continuous manner as the density variation of noise increases from its threshold.

9. The method according to claim 1, wherein said subdividing process comprises multiplication of said noise component by random numbers.

10. The method according to claim 1, wherein said sharpness enhancing process is Gaussian unsharp masking.

11. The method according to claim 1, wherein said smoothing process is Gaussian blurry masking.

12. The method according to claim 1, wherein said noise is graininess in the digital image.

13. The method according to claim 1, wherein said subdividing process is one of subdividing noise mottles.

14. An apparatus for processing digital images to suppress their noise and enhance their sharpness, comprising:

sharpness enhancing means for generating sharpness enhanced image data from original image data;

smoothing means for generating smoothed image data from said original image data;

extracting means for extracting a component containing both edges and noise from said sharpness enhanced image data and the resulting smoothed image data;

weighting data computing means for calculating weighting data for the edge and noise regions of said original image date;

separating means for separating a noise component from said edge/noise containing component;

subdividing means for producing a subdivided noise component from the separated noise component;

means of weighting by the weighting data for said edge region so as to provide an edge enhancing component from said edge/noise containing component; and

processed image data computing means for calculating processed image data from said sharpness enhancing image data, said subdivided noise component and said edge enhancing component.

15. The apparatus according to claim 14, wherein said weighting data computing means has edge detecting means for performing direct edge detection from said original image data and calculates the weighting data for said edge and noise regions by an edge component as detected by said edge detecting means.

16. The apparatus according to claim 15, wherein said separating means separates said noise component by performing nonlinear transformation after said edge/noise containing component extracted by said extracting means is weighted by the weighting data for said noise region as calculated by said weighting data computing means.

17. The apparatus according to claim 14, wherein said separating means separates an edge component and said noise component by directly performing nonlinear transformation on said edge/noise containing component as extracted by said extracting means.

18. The apparatus according to claim 17, wherein said weighting data computing means calculates the weighting data for said edge and noise regions from said edge component as separated by said separating means.