APPARATUS AND METHOD FOR ESTIMATING NOISE

• US 20140071156A1
• Filed: 09/11/2013
• Published: 03/13/2014
• Est. Priority Date: 09/11/2012
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: LEE, Hyun-seung, KIM, Seung-gu, PARK, Se-Hyeok
• Granted Patent: US 9,305,336 B2
• Time in Patent Office: Days
• US Class Current: 345/611
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/30168 Image quality inspection
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06T 7/0002 Inspection of images, e.g. flaw detection

Abstract

A noise estimation apparatus and method to prevent texture from being erroneously determined as noise, is provided. To this end, the noise estimation method includes generating an initial noise map for an input image signal, generating at least one noise map based on a temporal change of an image signal continuing from the image signal, calculating a correlation value for each region between the initial noise map and the noise maps, and determining noise of an image according to a temporal change based on the correlation value.

Claims

1. A noise estimation method comprising:

generating an initial noise map for an input image signal;

generating at least one noise map based on a first temporal change of an image signal continuing from the input image signal;

calculating a correlation value for each region between the initial noise map and the at least one noise map; and

determining noise of an image according to a second temporal change based on the correlation value.

2. The noise estimation method of claim 1, further comprising adjusting a strength of noise for each region based on a result of the noise determination.

3. The noise estimation method of claim 1, wherein the initial noise map is determined by:

Map_n=Y_n−

Y_NR,n,

where Map_n indicates a noise map of an n^th frame, and Y_NR,n indicates a noise-cancellation result of the n^th frame, and Y_n indicates an image which has not yet been noise-canceled.

4. The noise estimation method of claim 1, wherein the correlation value is determined by:

5. The noise estimation method of claim 1, wherein the determining of the noise of the image comprises determining a final noise dispersion map based on the correlation value, and the final noise dispersion map is determined by:

6. The noise estimation method of claim 1, wherein the determining of the noise of the image comprises:

comparing the correlation value with a threshold value;

excluding a corresponding portion from the noise determination if the correlation value is less than the threshold value; and

determining the corresponding portion as noise if the correlation value is greater than the threshold value.

13. The noise estimation method of claim 4, wherein the constant value is changeable.

14. The noise estimation method of claim 5, wherein the constant value is changeable.

15. The noise estimation method of claim 6, wherein the threshold value is predetermined.

7. A noise estimation apparatus comprising:

a noise map generator configured to generate an initial noise map for an input image signal and generate at least one noise map based on a first temporal change of an image signal continuing from the input image signal; and

a noise estimator configured to calculate a correlation value for each region between the initial noise map and the at least one noise map and determine noise of an image according to a second temporal change based on the correlation value.

8. The noise estimation apparatus of claim 7, wherein the noise estimator adjusts a strength of noise for each region based on a result of the noise determination.

9. The noise estimation apparatus of claim 7, wherein the initial noise map is determined by:

Map_n=Y_n−

Y_NR,n,

where Map_n indicates a noise map of an n^th frame, and Y_NR,n indicates a noise-cancellation result of the n^th frame, and Y_n indicates an image which has not yet been noise-canceled.

10. The noise estimation apparatus of claim 7, wherein the correlation value is determined by:

11. The noise estimation apparatus of claim 7, wherein the noise estimator determines a final noise dispersion map based on the correlation value, and the final noise dispersion map is determined by:

12. The noise estimation apparatus of claim 7, wherein the noise estimator, when determining the noise of the image, compares the correlation value with a threshold value,

excludes a corresponding portion from the noise determination if the correlation value is less than the threshold value, and determines the corresponding portion as noise if the correlation value is greater than the threshold value.

16. The noise estimation apparatus of claim 10, wherein the constant value is changeable.

17. The noise estimation apparatus of claim 11, wherein the constant value is changeable.

18. The noise estimation apparatus of claim 12, wherein the threshold value is predetermined.