System and method for reducing noise in an image

• US 8,472,724 B2
• Filed: 06/30/2010
• Issued: 06/25/2013
• Est. Priority Date: 06/30/2010
• Status: Expired due to Fees

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Lertrattanapanich, Surapong, Kim, Yeong-Taeg
• Primary Examiner(s): Allison, Andrae S
• Application Number:
  US12/827,986
• Publication Number:
  US 20120002896A1
• Time in Patent Office: 1,091 Days
• Field of Search: 382/199, 382/165, 382/167, 382/260, 382/261, 382/264, 382/266, 382/275, 348/606, 348/607, 348/622, 348/625, 348/627, 348/666, 375/240.2, 375/240.24, 375/240.29
• US Class Current: 382/205
• CPC Class Codes: G06T 2207/10004 Still image; Photographic image
• CPC Class Codes: G06T 5/70 Denoising; Smoothing

Abstract

A system and method of reducing noise are disclosed. In one embodiment a system comprises an input configured to receive an input image, a filter configured to filter the input image to generate a filtered image, a weight generator configured to generate one or more edge weights related to at least one edge of the input image, a mixer configured to generate an output image based on the input image, the filtered image, and at least one of the edge weights, and an output configured to output the output image.

Claims

1. A method of processing an image, the method comprising:

receiving an input image;

filtering the input image to generate a filtered image including determining a filtered value for each of the plurality of pixel locations by;

determining, for each of the plurality of pixel locations within a defined window about the particular pixel location, a difference between the input value associated with the particular location and the input value associated with the pixel locations within the defined window

determining, for each of the plurality of pixel locations within the defined window, a filter weight based on the difference between the input value associated with the particular pixel location and the input value associated with the pixel location within the defined window includes the filter weight based on a first difference greater than the filter weight based on a second difference where the first difference is less than the second difference,

determining, for each of the filter weights, a modified filter weight based on the filter weight and a preset window weight, and

determining the filtered value based on the differences and the modified filter weights;

generating one or more edge weights related to at least one edge of the input image;

generating an output image based on the input image, the filtered image, and at least one of the edge weights; and

outputting the output image.

2. The method of claim 1, wherein filtering the input image comprises sigma filtering the input image.

3. The method of claim 1, wherein generating an output image comprises mixing the input image and the filtered image based on the edge weights.

4. The method of claim 1, wherein:

receiving the input image comprises receiving a plurality of input values respectively associated with a plurality of pixel locations;

generating one or more edge weights comprises determining, for each of the plurality of pixel locations, an edge weight;

generating the output image comprises determining, for each of the plurality of pixel locations, an output value based on the input value associated with the pixel location, the filtered value determined for the pixel location, and the edge weight determined for the pixel location; and

outputting the output image comprises outputting the plurality of output values determined for each of the plurality of pixel locations respectively associated with the plurality of pixel locations.

5. The method of claim 4, wherein determining the edge weight for a particular pixel location comprises:

determining a first sigma weight based on a variation of the input values associated with a plurality of pixel locations within a defined first window about the particular pixel location;

determining a second sigma weight based on a variation of the input values associated with a plurality of pixel locations within a defined second window about

the particular pixel location, wherein the second window is larger than the first window; and

determining the edge weight based on the first sigma weight and the second sigma weight.

6. The method of claim 5, wherein the first sigma weight is determined such that, where a first variation is less than a second variation, a first sigma weight based on the first variation is greater than a first sigma weight based on the second variation.

7. The method of claim 5, wherein the second sigma weight is determined such that, where a first variation is less than a second variation, a second sigma weight based on the first variation is less than a second sigma weight based on the second variation.

8. The method of claim 5, wherein determining the edge weight comprises determining the edge weight based on the product of the first sigma weight and the second sigma weight.

9. The method of claim 4, wherein determining the output value for a particular pixel location comprises:

adding the input value associated with the particular pixel location and the filtered value determined for the particular pixel location, wherein at least one of the input value associated with the particular pixel location and the filtered value determined for the particular pixel location is weighted by the edge weight determined for the particular pixel location.

10. The method of claim 1, wherein the preset window weights for pixel locations within the defined window and closer to the particular pixel location are greater than preset window weights for pixel locations within the defined window and further from the particular pixel location.

11. The method of claim 1, wherein determining the filtered values comprises determining the filtered values based on the ratio of a sum of the differences weighted by the modified filter weights and a sum of the modified filter weights.

12. A system for processing an image, the system comprising:

a processor coupled to a memory;

a input configured to receive an input image;

a filter configured to;

determine a filtered value for each of a plurality of pixel locations,

determine, for a particular pixel location for each of the plurality of pixel locations within a defined window about the particular pixel location, a difference between an input value associated with the particular location and the input value associated with the pixel locations within the defined window,

determine, for each of the plurality of pixel locations within the defined window, a filter weight based on the difference between the input value associated with the particular pixel location and the input value associated with the pixel location within the defined window includes the filter weight based on a first difference greater than the filter weight based on a second difference where the first difference is less than the second difference,

determine, for each of the filter weights, a modified filter weight based on the filter weight and a preset window weight,

determine the filtered value based on the differences and the modified filter weights, and

generate a filtered image from the input image;

a weight generator configured to generate one or more edge weights related to at least one edge of the input image;

a mixer configured to generate an output image based on the input image, the filtered image, and at least one of the edge weights; and

an output configured to output the output image.

13. The system of claim 12, wherein the filter comprises a sigma filter.

14. The system of claim 12, wherein the mixer is configured to mix the input image and the filtered image based on the edge weights.

15. The system of claim 12, wherein:

the input is configured to receive a plurality of input values respectively associated with a plurality of pixel locations;

the weight generator is configured to determine, for each of the plurality of pixel locations, an edge weight;

the mixer is configured to determine, for each of the plurality of pixel locations, a output value based on the input value associated with the pixel location, the filtered value determined for the pixel location, and the weighting value determined for the pixel location; and

the output is configured to output the plurality of output values determined for each of the plurality of pixel locations respectively associated with the plurality of pixel locations.

16. The system of claim 15, wherein the weight generator is configured to, for a particular pixel location:

determine a first sigma weight based on a variation of the input values associated with a plurality of pixel locations within a defined first window about the particular pixel location;

determine a second sigma weight based on a variation of the input values associated with a plurality of pixel locations within a defined second window about the particular pixel location, wherein the second window is larger than the first window; and

determine the edge weight based on the first sigma weight and the second sigma weight.

17. The system of claim 16, wherein the weight generator is configured to determine the first sigma weight such that, where a first variation is less than a second variation, a first sigma weight based on the first variation is greater than a first sigma weight based on the second variation.

18. The system of claim 16, wherein the weight generator is configured to determine the second sigma weight such that, where a first variation is less than a second variation, a second sigma weight based on the first variation is less than a second sigma weight based on the second variation.

19. The system of claim 16, wherein the weight generator is configured to determine the edge weight based on the product of the first sigma weight and the second sigma weight.

20. The system of claim 15, wherein the mixer is configured to, for a particular pixel location:

add the input value associated with the particular pixel location and the filtered value determined for the particular pixel location, wherein at least one of the input value associated with the particular pixel location and the filtered value determined for the particular pixel location is weighted by the edge weight determined for the particular pixel location.

21. The system of claim 12, wherein the preset window weights for pixel locations within the defined window and closer to the particular pixel location are greater than preset window weights for pixel locations within the defined window and further from the particular pixel location.

22. The system of claim 12, wherein the filter is configured to determine the filtered values based on the ratio of a sum of the differences weighted by the modified filter weights and a sum of the modified filter weights.

23. A system for processing an image, the system comprising:

means for receiving an input image;

means for filtering configured to;

determine a filtered value for each of a plurality of pixel locations,

determine, for a particular pixel location for each of the plurality of pixel locations within a defined window about the particular pixel location, a difference between an input value associated with the particular location and the input value associated with the pixel locations within the defined window,

determine, for each of the plurality of pixel locations within the defined window, a filter weight based on the difference between the input value associated with the particular pixel location and the input value associated with the pixel location within the defined window includes the filter weight based on a first difference greater than a filter weight based on a second difference where the first difference is less than the second difference,

determine, for each of the filter weights, a modified filter weight based on the filter weight and a preset window weight,

determine the filtered value based on the differences and the modified filter weights, and

generate a filtered image from the input image;

means for generating one or more edge weights related to at least one edge of the input image;

means for generating an output image based on the input image, the filtered image, and at least one of the edge weights; and

means for outputting the output image.

24. The system of claim 23, wherein:

the means for receiving the input image comprises means for receiving a plurality of input values respectively associated with a plurality of pixel locations;

the means for generating one or more edge weights comprises means for determining, for each of the plurality of pixel locations, an edge weight;

the means for generating an output image comprises means for determining, for each of the plurality of pixel locations, an output value based on the input value associated with the pixel location, the filtered value determined for the pixel location, and the edge weight determined for the pixel location; and

the means for outputting the output image comprises means for outputting the plurality of output values determined for each of the plurality of pixel locations respectively associated with the plurality of pixel locations.

25. The system of claim 23, wherein at least one of the means for receiving, means for filtering, means for generating one or more edge weights, means for generating an output image, and means for outputting comprises a processor.

26. A non-transitory computer-readable storage medium having processor-executable instructions encoded thereon which, when executed by a processor, cause a computer to perform a method of image processing, the method comprising:

receiving an input image;

filtering the input image to generate a filtered image including determining a filtered value for each of the plurality of pixel locations by;

determining, for each of the plurality of pixel locations within a defined window about the particular pixel location, a difference between the input value associated with the particular location and the input value associated with the pixel locations within the defined window

determining, for each of the plurality of pixel locations within the defined window, a filter weight based on the difference between the input value associated with the particular pixel location and the input value associated with the pixel location within the defined window includes the filter weight based on a first difference greater than the filter weight based on a second difference where the first difference is less than the second difference,

determining, for each of the filter weights, a modified filter weight based on the filter weight and a preset window weight, and

determining the filtered value based on the differences and the modified filter weights;

generating one or more edge weights related to at least one edge of the input image;

generating an output image based on the input image, the filtered image, and at least one of the edge weights; and

outputting the output image.

27. The non-transitory computer-readable storage medium of claim 26, wherein:

receiving the input image comprises receiving a plurality of input values respectively associated with a plurality of pixel locations;

generating one or more edge weights comprises determining, for each of the plurality of pixel locations, a edge weight;

generating the output image comprises determining, for each of the plurality of pixel locations, an output value based on the input value associated with the pixel location, the filtered value determined for the pixel location, and the edge weight determined for the pixel location; and

outputting the output image comprises outputting the plurality of output values determined for each of the plurality of pixel locations respectively associated with the plurality of pixel locations.