ADAPTIVE FILTER FOR LOW-LIGHT NOISE REDUCTION

• US 20120127370A1
• Filed: 11/19/2010
• Published: 05/24/2012
• Est. Priority Date: 11/19/2010
• Status: Active Grant

• Current Assignee: Analog Devices, Inc.
• Original Assignee: Analog Devices, Inc.
• Inventors: SINGH, RAKA, Malik, Gaurav, Mahapatra, Rajesh
• Granted Patent: US 8,699,813 B2
• Time in Patent Office: Days
• US Class Current: 348/607
• CPC Class Codes: H04N 5/213 Circuitry for suppressing or minimising impulsive noise for suppressing or minimising disturbance in image signal generation H04N23/81

Abstract

In general, in one embodiment, low-light noise is removed from an image by separately filtering luma and chroma components of the image, by adaptively filtering the image based at least in part on a Gaussian distribution of the image, and/or by dividing the image into separate regions and filtering each region separately.

Claims

1. An adaptive filter for filtering noise from a low-light image, the adaptive filter comprising:

a morphology filter for dilating an edge detected in the image, thereby dividing the image into a non-edge region and a dilated edge region; and

a comparative filter for comparing a pixel in the non-edge region to a region surrounding the pixel and optionally replacing the pixel, based on a result of the comparison, with a new pixel derived at least in part from the region surrounding the pixel.

2. The adaptive filter of claim 1, further comprising a difference filter for detecting the edge in the image.

3. The adaptive filter of claim 1, wherein the morphology filter comprises a dilation filter.

4. The adaptive filter of claim 3, wherein the dilation filter is a 3×

4 dilation filter.

5. The adaptive filter of claim 1, wherein the region surrounding the pixel corresponds to a 3×

3 pixel region.

6. The adaptive filter of claim 1, wherein the adaptive filter compares a difference between the pixel and a mean of the region surrounding the pixel to a variance of the region surrounding the pixel.

7. The adaptive filter of claim 6, wherein the variance is a variance of a Gaussian distribution of the region surrounding the pixel.

8. The adaptive filter of claim 7, further comprising a circuit for computing the variance of the Gaussian distribution of the region surrounding the pixel.

9. The adaptive filter of claim 6, wherein the mean is a mean of a Gaussian distribution of the region surrounding the pixel.

10. The adaptive filter of claim 7, further comprising a circuit for computing the mean of the Gaussian distribution of the region surrounding the pixel.

11. The adaptive filter of claim 1, wherein the comparative filter derives the new pixel based at least in part on a median value of the region surrounding the pixel.

12. The adaptive filter of claim 1, wherein the comparative filter derives the new pixel based at least in part on a low-pass filter value of the region surrounding the pixel.

13. The adaptive filter of claim 1, wherein the comparative filter replaces a second pixel adjacent to the pixel with the new pixel.

14. The adaptive filter of claim 13, wherein the pixel is an even pixel and the second pixel is an odd pixel.

15. The adaptive filter of claim 1, wherein the comparative filter preserves the pixel.

16. A method for adaptively filtering noise from a low-light image, the method comprising:

dilating an edge detected in an image to divide the image into a non-edge region and a dilated edge region;

comparing a pixel in the non-edge region to a region surrounding the pixel; and

optionally replacing the pixel, based on a result of the comparison, with a new pixel derived at least in part from the region surrounding the pixel.

17. The method of claim 16, further comprising detecting the edge in the image.

18. The method of claim 16, wherein comparing the pixel to the region surrounding the pixel comprises comparing a difference between the pixel and a variance of the region surrounding the pixel to a mean of the region surrounding the pixel.

19. The method of claim 18, wherein the variance is a variance of a Gaussian distribution of the region surrounding the pixel.

20. The method of claim 18, wherein the mean is a mean of a Gaussian distribution of the region surrounding the pixel.

21. The method of claim 16, wherein the new pixel is derived based at least in part on a median value of the region surrounding the pixel.

22. The method of claim 16, wherein the new pixel is derived based at least in part on a low-pass filter value of the region surrounding the pixel.

23. The method of claim 16, further comprising replacing a second pixel with the new pixel, the second pixel being adjacent to the pixel.

24. The method of claim 23, wherein the pixel is an even pixel and the second pixel is an odd pixel.

25. The method of claim 16, wherein the comparative filter preserves the pixel.