Noise removal device, noise removal method, and computer readable recording medium

• US 8,391,635 B2
• Filed: 08/28/2008
• Issued: 03/05/2013
• Est. Priority Date: 08/31/2007
• Status: Expired due to Fees

• Current Assignee: Kanazawa University, Olympus Corporation
• Original Assignee: Kanazawa University, Olympus Corporation
• Inventors: Aragaki, Hideya, Saito, Takahiro, Komatsu, Takashi
• Primary Examiner(s): Bella, Matthew
• Assistant Examiner(s): Thirugnanam, Gandhi
• Application Number:
  US12/200,243
• Publication Number:
  US 20090060323A1
• Time in Patent Office: 1,650 Days
• Field of Search: 382/164, 382/173
• US Class Current: 382/275
• CPC Class Codes: G06T 2207/20064 Wavelet transform [DWT]
• CPC Class Codes: G06T 5/10 using non-spatial domain filtering
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06V 10/30 Noise filtering

Abstract

A noise removal device that removes noise from an image signal includes: a separation unit that separates the image signal into at least two subband signals; a parameter setting unit that sets a plurality of coring ranges in relation to at least one of the subband signals; and a coring unit that performs coring processing on the basis of the plurality of coring ranges.

Claims

1. A noise removal device that removes noise from an image signal, comprising:

a separation unit that separates the image signal into at least two subband signals;

a noise estimation unit that estimates a noise amount of a pixel of at least one of the subband signals;

a parameter setting unit that sets a plurality of coring ranges in relation to at least one of the subband signals, the coring ranges being ranges of signal values of subband signals to be subjected to coring processing within the subband signals; and

a coring unit that performs the coring processing on the at least one of the subband signals based on the plurality of coring ranges;

wherein the parameter setting unit includes a color difference/color sum calculation unit that calculates at least one of a color difference and a color sum using signal values in a region within the at least one of the subband signals;

wherein the parameter setting unit adjusts the plurality of coring ranges based on the noise amount and the at least one of the color difference and the color sum;

wherein the parameter setting unit sets a first fixed output value for a first coring range in the plurality of coring ranges in relation to the at least one of the subband signals, and sets a second fixed output value for a second coring range in the plurality of coring ranges in relation to the at least one of the subband signals; and

wherein the coring unit;

transforms signal values belonging to the first coring range within the at least one of the subband signals into the first fixed output value,

transforms signal values belonging to the second coring range within the at least one of the subband signals into the second fixed output value, and

corrects signal values that do not belong to any one of the plurality of coring ranges in relation to the at least one of the subband signals.

2. The noise removal device as defined in claim 1, wherein the image signal is comprised of a plurality of color components, and operations of the separation unit, the parameter setting unit, and the coring unit are applied to each of the plurality of color components.

3. The noise removal device as defined in claim 2, wherein the parameter setting unit sets the plurality of coring ranges based on a calculation value of a signal value of the plurality of color components.

4. The noise removal device as defined in claim 1, wherein the separation unit employs a wavelet transform.

5. The noise removal device as defined in claim 1, further comprising a synthesis unit that synthesizes the plurality of subband signals following the coring processing.

6. A noise removal method for removing noise from an image signal, the method comprising:

separating the image signal into at least two subband signals;

estimating a noise amount of a pixel of at least one of the subband signals;

setting a plurality of coring ranges in relation to at least one of the subband signals, the coring ranges being ranges of signal values of subband signals to be subjected to coring processing within the subband signals; and

performing the coring processing on the at least one of the subband signals based on the plurality of coring ranges;

wherein the setting comprises;

calculating at least one of a color difference and a color sum using signal values in a region within the at least one of the subband signals;

adjusting the plurality of coring ranges based on the noise amount and the at least one of the color difference and the color sum; and

setting (i) a first fixed output value for a first coring range in the plurality of coring ranges in relation to the at least one of the subband signals, and (ii) a second fixed output value for a second coring range in the plurality of coring ranges in relation to the at least one of the subband signals; and

wherein the coring comprises;

transforming signal values belonging to the first coring range within the at least one of the subband signals into the first fixed output value,

transforming signal values belonging to the second coring range within the at least one of the subband signals into the second fixed output value, and

correcting signal values that do not belong to any one of the plurality of coring ranges in relation to the at least one of the subband signals.

7. The noise removal method as defined in claim 6, wherein the image signal is comprised of a plurality of color components, and the separating, the setting, and the coring are applied to each of the plurality of color components.

8. The noise removal method as defined in claim 7, wherein the setting comprises setting the plurality of coring ranges based on a calculation value of a signal value of the plurality of color components.

9. The noise removal method as defined in claim 6, wherein the separating employs a wavelet transform.

10. The noise removal method as defined in claim 6, further comprising synthesizing the plurality of subband signals after the coring is performed.

11. A non-transitory computer readable storage medium having a program stored thereon, the program causing a computer to execute a method for removing noise from an image signal, the method comprising:

separating the image signal into at least two subband signals;

estimating a noise amount of a pixel of at least one of the subband signals;

setting a plurality of coring ranges in relation to at least one of the subband signals, the coring ranges being ranges of signal values of subband signals to be subjected to coring processing within the subband signals; and

performing the coring processing on the at least one of the subband signals based on the plurality of coring ranges;

wherein the setting comprises;

calculating at least one of a color difference and a color sum using signal values in a region within the at least one of the subband signals;

adjusting the plurality of coring ranges based on the noise amount and the at least one of the color difference and the color sum; and

setting (i) a first fixed output value for a first coring range in the plurality of coring ranges in relation to the at least one of the subband signals, and (ii) a second fixed output value for a second coring range in the plurality of coring ranges in relation to the at least one of the subband signals; and

wherein the coring comprises;

transforming signal values belonging to the first coring range within the at least one of the subband signals into the first fixed output value,

transforming signal values belonging to the second coring range within the at least one of the subband signals into the second fixed output value, and

correcting signal values that do not belong to any one of the plurality of coring ranges in relation to the at least one of the subband signals.