KR101707337B1 - Multiresolution non-local means filtering method for image denoising - Google Patents

Multiresolution non-local means filtering method for image denoising Download PDF

Info

Publication number
KR101707337B1
KR101707337B1 KR1020150165299A KR20150165299A KR101707337B1 KR 101707337 B1 KR101707337 B1 KR 101707337B1 KR 1020150165299 A KR1020150165299 A KR 1020150165299A KR 20150165299 A KR20150165299 A KR 20150165299A KR 101707337 B1 KR101707337 B1 KR 101707337B1
Authority
KR
South Korea
Prior art keywords
signal
noise
frequency band
image
nlm
Prior art date
Application number
KR1020150165299A
Other languages
Korean (ko)
Inventor
권구락
Original Assignee
조선대학교산학협력단
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 조선대학교산학협력단 filed Critical 조선대학교산학협력단
Priority to KR1020150165299A priority Critical patent/KR101707337B1/en
Application granted granted Critical
Publication of KR101707337B1 publication Critical patent/KR101707337B1/en

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/05—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves 
    • A61B5/055—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves  involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/0033—Features or image-related aspects of imaging apparatus classified in A61B5/00, e.g. for MRI, optical tomography or impedance tomography apparatus; arrangements of imaging apparatus in a room
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/725—Details of waveform analysis using specific filters therefor, e.g. Kalman or adaptive filters
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R33/00—Arrangements or instruments for measuring magnetic variables
    • G01R33/20—Arrangements or instruments for measuring magnetic variables involving magnetic resonance
    • G01R33/28—Details of apparatus provided for in groups G01R33/44 - G01R33/64
    • G01R33/32—Excitation or detection systems, e.g. using radio frequency signals
    • G01R33/36—Electrical details, e.g. matching or coupling of the coil to the receiver
    • G01R33/3607—RF waveform generators, e.g. frequency generators, amplitude-, frequency- or phase modulators or shifters, pulse programmers, digital to analog converters for the RF signal, means for filtering or attenuating of the RF signal
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R33/00—Arrangements or instruments for measuring magnetic variables
    • G01R33/20—Arrangements or instruments for measuring magnetic variables involving magnetic resonance
    • G01R33/44—Arrangements or instruments for measuring magnetic variables involving magnetic resonance using nuclear magnetic resonance [NMR]
    • G01R33/48—NMR imaging systems
    • G01R33/54—Signal processing systems, e.g. using pulse sequences ; Generation or control of pulse sequences; Operator console
    • G01R33/56—Image enhancement or correction, e.g. subtraction or averaging techniques, e.g. improvement of signal-to-noise ratio and resolution
    • G01R33/5608—Data processing and visualization specially adapted for MR, e.g. for feature analysis and pattern recognition on the basis of measured MR data, segmentation of measured MR data, edge contour detection on the basis of measured MR data, for enhancing measured MR data in terms of signal-to-noise ratio by means of noise filtering or apodization, for enhancing measured MR data in terms of resolution by means for deblurring, windowing, zero filling, or generation of gray-scaled images, colour-coded images or images displaying vectors instead of pixels

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Surgery (AREA)
  • Veterinary Medicine (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Signal Processing (AREA)
  • Animal Behavior & Ethology (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Public Health (AREA)
  • Pathology (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Radiology & Medical Imaging (AREA)
  • High Energy & Nuclear Physics (AREA)
  • Condensed Matter Physics & Semiconductors (AREA)
  • General Physics & Mathematics (AREA)
  • Physiology (AREA)
  • Psychiatry (AREA)
  • Magnetic Resonance Imaging Apparatus (AREA)

Abstract

The present invention relates to a multi-resolution NLM filtering method for image degusting for decomposing a wavelet decomposition signal and a frequency into subbands and then applying the restored signal to an approximate subband after NLM filtering to remove low frequency noise components of the MRI image , Applying a wavelet transform to cause noise images to appear in multiple subbands, allowing the NLM filter to be applied to low-frequency subbands of signal decomposition using a wavelet filter bank, And combining the wavelet thresholds to fabricate a gonioscopic image framework and effectively remove the noise included in the MR image. Therefore, it is possible to provide an MRI image having better quality by removing noise from a MRI image including fixed noise in a shorter time than a conventional NLM method, and maintaining the characteristics of the MRI image.

Figure 112015114965714-pat00010

Description

[0002] Multiresolution non-local means filtering method for image denoising [

The present invention relates to a multi-resolution NLM filtering method for image gouging, in particular, decomposes a wavelet decomposition signal and a frequency into subbands, NLM-filters the reconstructed signal, and applies it to an approximate subband, Resolution NLM filtering method for image ghosting to remove components.

Noise rejection in images has been studied for various applications for over half a century. The basic purpose of noise reduction in an image is to effectively reduce noise while maintaining features. For use in medical applications, magnetic resonance imaging (MRI) eliminates phase information to prevent artificial phase formation. MR images are sensitive to artificial machining or noise sources.

The noise of the MR (magnetic resonance) signal is caused by a change in the severity of the deterioration through thermal oscillation of ions and electrons in the receiving coil and the sample in useful clinical image information of a part of the MR image.

A number of noise reduction methods have been developed over the years. Among these methods, the method of using the threshold value of wavelets is the most widely used method. It is an excellent localization property of wavelet and is used as an essential signal processing and image processing tool for various applications.

Wavelength doughening removes the noise in it while maintaining the characteristics of the signal. In the wavelet threshold, the signal is decomposed into low frequency and high frequency subbands, and since most of the image information is concentrated on several large coefficients, high frequency subbands are processed through hard or soft threshold operations.

The most important task in wavelet critical processing is the selection of threshold values. Various threshold selection strategies are proposed, for example VisuShrink, SureShrink and BayesShrink.

In the VisuShrin scheme, the universal threshold is developed by a function of the deviation of the noise and the number of samples based on the minimum and maximum error coefficients. SureShrink thresholds are optimized in terms of Stein's equitable risk estimates. The BayesShrink method determines thresholds in the Bayesian framework through Gaussian distribution modeling of wavelet coefficients.

In addition, Wang developed a Haar-based moving window that performs two wavelet transforms for local multi-scale analysis. In this way, the wavelet transform is performed in a moving window. DWT was used to remove noise from successful images.

However, there has been a demand for development of a better noise estimation method in order to provide an MRI image in which noise is effectively removed.

Korean Patent No. 10-1310825

In order to satisfy such a development requirement, the present invention is a method for decomposing a wavelet decomposition signal and a frequency into subbands, applying NLM filtering to the reconstructed signal and then applying the wavelet decomposition signal and the reconstructed signal to an approximate subband to remove a low- Resolution NLM filtering method for digital signal processing.

According to an embodiment of the present invention, there is provided a multi-resolution NLM filtering method for image gaining, comprising: inputting a photographed MR (Magnetic Resonance) image; Dividing the input MR image into an MR signal of a low frequency band and an MR signal of a high frequency band; Frequency-band signal and the high-frequency-band signal to generate one approximate value and one wavelet coefficient, and dividing the divided high-frequency-band MR signal into a low-frequency-band signal and a high- To generate two wavelet coefficients; By applying the approximate value to the NLM (Non-Local Means) filter, applying non-local mean (NLM) filtering and applying the soft threshold value to the three wavelet coefficients, the wavelet coefficients are adjusted to remove noise ; A low frequency band signal is generated by summing the NLM filtered low frequency band signal and the noise canceled high frequency band signal by the wavelet coefficient and the low frequency band signal and the high frequency band signal from which the noise is removed by the wavelet coefficient are summed Generating a signal of a high frequency band; And a step of summing signals of a low-frequency band and a high-frequency band from which noise has been removed to restore the MR signal, thereby providing a noise-free MR image.

As an embodiment related to the present invention, the soft threshold values can be applied to the real and imaginary parts of the wavelet coefficients, respectively, to adjust the wavelet coefficients.

As an embodiment related to the present invention, the soft threshold value is selected based on the value calculated by multiplying the standard deviation of the noise by a constant value (TM), and the noise can be derived based on an average of the standard deviation.

The present invention decomposes a wavelet decomposition signal and a frequency into subbands, and NLM-filtered the reconstructed signal and then applies it to an approximate subband to remove a low-frequency noise component of the MRI image. It is effective to remove noise in a shorter time than the NLM method, to maintain the characteristics of the MRI image, and to provide an MRI image having a better quality.

FIG. 1 is a view for explaining a configuration of a multi-resolution NLM filtering system for video gnoving according to the present invention.
FIG. 2 is a block diagram for explaining a multi-resolution NLM filtering method for video gnoving according to the present invention.
FIG. 3 is a diagram showing a filtering result of an MR image applied to the present invention.

It is noted that the technical terms used in the present invention are used only to describe specific embodiments and are not intended to limit the present invention. In addition, the technical terms used in the present invention should be construed in a sense generally understood by a person having ordinary skill in the art to which the present invention belongs, unless otherwise defined in the present invention, Should not be construed to mean, or be interpreted in an excessively reduced sense. In addition, when a technical term used in the present invention is an erroneous technical term that does not accurately express the concept of the present invention, it should be understood that technical terms can be understood by those skilled in the art. In addition, the general terms used in the present invention should be interpreted according to a predefined or prior context, and should not be construed as being excessively reduced.

Furthermore, the singular expressions used in the present invention include plural expressions unless the context clearly dictates otherwise. In the present invention, terms such as "comprising" or "comprising" and the like should not be construed as encompassing various elements or stages of the invention, Or may further include additional components or steps.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings, wherein like reference numerals refer to like or similar elements throughout the several views, and redundant description thereof will be omitted.

FIG. 1 is a view for explaining a configuration of a multi-resolution NLM filtering system for video gnoving according to the present invention.

1, a multi-resolution NLM filtering system 100 for video gnoving applied to the present invention includes an input unit 110, a control unit 120, a photographing unit 130, a noise processing unit 140, (150).

The input unit 110 outputs a multi-resolution NLM filtering routine execution request signal in response to a physical force externally applied thereto and a photographing request signal output to the photographing unit 130. The multi-resolution NLM filtering system 100 And provides a variety of menus to be provided by the user.

The control unit 120 allows the corresponding menus to be provided through the input unit 110 as a user interface when the multi-resolution NLM filtering system 100 for image dynaming is operated, and displays the menu in response to a photographing request signal input through the input unit 110. [ The MRI image taken through the imaging unit 130 is output to the noise processing unit 140 so that the MRI image with the noise removed is output through the output unit 150 And controls the system.

The noise processing unit 140 receives an MR (Magnetic Resonance) image photographed through the imaging unit 130 and divides the input MR image into an MR signal of a low frequency band and an MR signal of a high frequency band, The MR signal is divided into a signal of a low frequency band and a signal of a high frequency band to generate one approximate value and one wavelet coefficient.

The noise processing unit 140 divides the divided high frequency band MR signal into a low frequency band signal and a high frequency band signal to generate two wavelet coefficients and applies an approximate value to an NLM (Non-Local Means) Non-Local Means) and applying soft thresholds to the three wavelet coefficients to adjust the wavelet coefficients to remove the noise with a small absolute value. The NLM-filtered low-frequency signals and the wavelet coefficients cause noise Frequency band signal by adding the removed high frequency band signals to generate a low frequency band signal and adding the low frequency band signal and the high frequency band signal from which noises have been removed by the wavelet coefficient to generate a high frequency band signal.

Then, the noise processing unit 140 restores the MR signal by summing the signals of the low-frequency band and the high-frequency band from which noise has been removed, and outputs the noise-removed MR image to the output unit 150 to be displayed on the screen .

Here, the soft threshold value is applied to the real part and the imaginary part of the wavelet coefficient to adjust the wavelet coefficient, respectively, and the soft threshold value is also selected based on the value calculated by multiplying the standard deviation of the noise by a constant value (TM) , The noise is derived based on an average of the standard deviation.

The multi-resolution NLM filtering method for the image dNOS is as follows.

FIG. 2 is a block diagram for explaining a multi-resolution NLM filtering method for video gnoving according to the present invention.

Figure 2 shows the block diagram of the proposed algorithm. Dinoizing  Multi-resolution for NLM Filtering  The method is described below. I have marked it in dotted line according to the order of progress.

(Included in the MR image noise  characteristic)

=> The proposed paper includes MR Noise  Is it the same meaning as described above?

Before describing the noise removal process of the MR image, the characteristics of the noise included in the MR image will be described first. The MRI is calculated in both real and virtual images, assuming the zero mean of the Gaussian noise distribution. Thus, the dependent noise of the image follows the Rician distribution, and noise removal is difficult. The complex MR image is expressed by the following equation (1).

Figure 112015114965714-pat00001

Where z is a complex MR image, Zreal and Zimaginary have real and imaginary components, and are independently impaired by n1 and n2 of white Gaussian noise, zero mean and standard deviation σ, r, The original size and phase. The noise of the MRI can be measured as shown in the following equation (2).

=> "Zero mean and standard deviation" for σ, r, to  It has been named. That is, there are three symbols, and the two names are limited, so the relationship may not be clear. Please give a supplementary explanation about this part.

Figure 112015114965714-pat00002

The size of the measured MRI is expressed in the Rician distribution as the following equation (3).

Figure 112015114965714-pat00003

Where I 0 denotes a Bessel function of zero in order. The noise of the measured magnitude of MRI is difficult to estimate. The most commonly used method is to square the size of the MRI.

(Included in the MR image noise  Removal routine)

2, the noise processing unit 140 may include a noise reduction unit 140 for removing noise included in the MR image and providing a clear MR image, ) Image (S100).

The noise image included in the input MR image is divided into several subbands in one level wavelet transform. That is, the noise processing unit 140 divides the input MR image into an MR signal of a low frequency band and an MR signal of a high frequency band (S120).

The noise processing unit 140 divides the divided MR signals of the low frequency band into a signal of a low frequency band and a signal of a high frequency band to generate one approximate value and one wavelet coefficient and outputs the MR signal of the divided high frequency band to a low frequency band And a high frequency band signal to generate two wavelet coefficients S (130).

That is, LL, HL, HH, and HL, which are decomposed into one approximate value and three wavelet coefficients, are shown in FIG. Where HH represents the approximation of horizontal and vertical wavelet coefficients in subband HL while providing diagonal wavelet coefficients.

=> Of the above description LL , HL , HH , HL In order  Additional explanation request . As seen in Figure 2, LL , LH , HL , HH  It can be said that it is written in order. And "here HH  diagonal Wavelet  While providing the coefficients, In HL  Horizontal and vertical Wavelet  Coefficient Approximate values  "Is it a known technology that everyone in the technical field knows? If so, please briefly mention the related materials.

The gradeless noise processing unit 140 applies NLM (Non-Local Means) filtering to an approximate value to a NLM (Non-Local Means) filter, adjusts wavelet coefficients by applying a soft threshold value to three wavelet coefficients, The small signal is removed to remove the noise (S140).

Suppose that X (n) is a damaged signal, the noise-free signal is S (n) and the noise is N (n). If the noise signal X (n) is wavelet-transformed, a wavelet coefficient can be obtained as shown in Equation (4).

=> "X (n) is damaged Signal called  When you do, Noise  The missing signal is S (n), Noi (N) To "  There are no mathematical expressions in which X (n), S (n), and N (n) are written.

Figure 112015114965714-pat00004

Y jk represents the coefficient of the noise image in the wavelet domain. u jk takes a noise-free coefficient, and v jk a noise coefficient. The wavelet shrink method is a general method for noise reduction in the wavelet domain. This method is based on an appropriate adjustment of the wavelet coefficient by the threshold value. This is an effective way to isolate noise and send signals using size reduction or motion. Small absolute values of the wavelet coefficients are considered to be mostly performing noise and less content in the signal. In contrast, a signal having a high value carries important information of a wavelet coefficient. Therefore, by removing the coefficients with small absolute values and restoring the signal, a signal with low noise is generated.

The above-mentioned soft threshold value is applied to the real part and the imaginary part of the wavelet coefficient and is expressed by the following equation (5).

Figure 112015114965714-pat00005

Where T denotes a threshold value, and u ' jk denotes inversely replaced coefficients.

The multi-resolution NLM filtering method for video gnozing according to the present invention can effectively select and calculate the threshold value and the threshold method. Various methods have been proposed for estimating the optimal threshold value, and most methods use white Gaussian noise and reduced wavelet transform. Fundamentally, this can be classified into two categories according to global threshold and threshold level .

The global threshold method selects one threshold T that can be applied to all wavelet coefficients. Different thresholds can be used between different levels in a level-dependent manner. In the present invention, the standard deviation of the noise is multiplied by a constant value TM. The threshold level can be described by Equation (6).

=> "Video" above Dinoizing  Multi-resolution for NLM Filtering  The method can be effectively calculated by selecting a threshold and a threshold method. "In the article," the algorithm effectively selects the threshold and the threshold method Calculate "  It is the part represented. . However,  It is difficult to pinpoint exactly where thresholds and critical methods are calculated effectively. Please explain more about this part.

Figure 112015114965714-pat00006

Where σ jk is the standard deviation of the noise of the k th scale. For adaptive estimation of the threshold level, the noise can be derived based on an average of the standard deviation.

=> Of the parts listed above, k th " You can not find content or mathematical expressions that contain. I would like to further explain this part.

Figure 112015114965714-pat00007

The noise processing unit 140 generates a signal of a low frequency band by summing the signal of the NLM filtered low frequency band and the signal of the high frequency band in which the noise is removed by the wavelet coefficient and outputs the signal of the low frequency band And the high frequency band signal to generate a high frequency band signal (S150).

Finally, the noise processor 140 sums the signals of the low-frequency band and the high-frequency band from which the noise is removed, restores the MR signal, and provides the MR image from which noise has been removed (S160).

As described above, the multi-resolution NLM filter means applied to the multi-resolution NLM filtering method according to the present invention has the potential to remove low-frequency noise components, unlike the standard single-level NLM filter.

When the gray scale image Y = {Y (i) | i? I} is applied to the present invention, the NLM filter is defined by the following equation (8).

Figure 112015114965714-pat00008

Where pixel i in NLY (i) is the weighted average of all gray values in the image, and c (x) is the normalization factor. The weight w (i, j) is defined by the following equation (9).

Figure 112015114965714-pat00009

The patch Y (N i ) and Y (N j ) of the vectorized image associated with the computation express the similarity between the neighbors of pixels N i and N j . h is used as a normalization factor. The standard deviation Gaussian kernel G (.) Is used to consider the distance between the center pixel of the patch and the other pixel.

FIG. 3 is a diagram showing a filtering result of an MR image applied to the present invention, wherein (a) is a original image, (b) is a noisy image, and (c) (D) is a filtered image applied to the present invention.

As shown in FIG. 3, the MR image obtained by the filtering operation according to the present invention and the NLM filtering are shown. Experimental results show that the generated ghost images of the proposed algorithm are not only subjective but also objectively better.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or essential characteristics thereof. Therefore, the embodiments disclosed in the present invention are intended to illustrate rather than limit the scope of the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The scope of protection of the present invention should be construed according to the following claims, and all technical ideas within the scope of equivalents should be construed as falling within the scope of the present invention.

100: Multi-resolution NLM filtering system
110: input unit 120:
130: photographing part 140: noise processing part
150:

Claims (3)

Receiving an MR (Magnetic Resonance) image photographed by the photographing unit;
Dividing the input MR image into an MR signal of a low frequency band and an MR signal of a high frequency band;
Frequency signal and a high-frequency band signal to generate one approximate value and one wavelet coefficient, and dividing the divided high-frequency-band MR signal into a low-frequency signal and a high- Dividing the signal into two wavelet coefficients;
The NLM (Non-Local Means) filter is applied to the approximate value by applying NLM (Non-Local Means) filter and the wavelet coefficients are adjusted by applying a soft threshold value to the three wavelet coefficients, Removing;
A low frequency band signal is generated by summing the NLM filtered low frequency band signal and the noise canceled high frequency band signal by the wavelet coefficient and the low frequency band signal and the high frequency band signal from which the noise is removed by the wavelet coefficient are summed Generating a signal of a high frequency band; And
Summing signals of the low-frequency band and the high-frequency band from which the noise is removed to recover an MR signal, thereby providing a noise-free MR image;
Resolution NLM filtering method for image ghosting.
The method according to claim 1,
Wherein the soft threshold value is applied to a real part and an imaginary part of the wavelet coefficient to adjust the wavelet coefficient, respectively.
The method according to claim 1,
The soft threshold value is selected based on a value calculated by multiplying a standard deviation of noise by a constant value (TM)
Wherein the noise is derived based on an average of standard deviation. ≪ RTI ID = 0.0 > 8. < / RTI >

KR1020150165299A 2015-11-25 2015-11-25 Multiresolution non-local means filtering method for image denoising KR101707337B1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1020150165299A KR101707337B1 (en) 2015-11-25 2015-11-25 Multiresolution non-local means filtering method for image denoising

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1020150165299A KR101707337B1 (en) 2015-11-25 2015-11-25 Multiresolution non-local means filtering method for image denoising

Publications (1)

Publication Number Publication Date
KR101707337B1 true KR101707337B1 (en) 2017-02-17

Family

ID=58265461

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1020150165299A KR101707337B1 (en) 2015-11-25 2015-11-25 Multiresolution non-local means filtering method for image denoising

Country Status (1)

Country Link
KR (1) KR101707337B1 (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106911410A (en) * 2017-05-02 2017-06-30 广东工业大学 One kind communication primary user's cognitive method and system
KR20190119766A (en) * 2018-04-13 2019-10-23 서울과학기술대학교 산학협력단 Method for detecting position of radiation source using rotating modulation collimator and imaging device for performing the method
CN115808648A (en) * 2022-11-18 2023-03-17 无锡鸣石峻致医疗科技有限公司 Device and method for measuring ringing noise of magnetic resonance system

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2012513816A (en) * 2008-12-25 2012-06-21 メディック ビジョン イメージング ソルーション リミティッド Noise removal from medical images
KR101310825B1 (en) 2012-05-10 2013-10-14 고려대학교 산학협력단 Method for generating a magnetic resonance imaging and apparatus for generating a magnetic resonance imaging thereof
JP2015104668A (en) * 2013-11-29 2015-06-08 株式会社日立製作所 Method and apparatus for removing artifact from image

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2012513816A (en) * 2008-12-25 2012-06-21 メディック ビジョン イメージング ソルーション リミティッド Noise removal from medical images
KR101310825B1 (en) 2012-05-10 2013-10-14 고려대학교 산학협력단 Method for generating a magnetic resonance imaging and apparatus for generating a magnetic resonance imaging thereof
JP2015104668A (en) * 2013-11-29 2015-06-08 株式会社日立製作所 Method and apparatus for removing artifact from image

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
Wavelet Based Non Local Means Algorithm for Efficient Denoising of MRI Images(International Journal of Advanced Research in Computer and Communication Engineering, Vol.4, Issue 2, 2015.02.) *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106911410A (en) * 2017-05-02 2017-06-30 广东工业大学 One kind communication primary user's cognitive method and system
CN106911410B (en) * 2017-05-02 2020-10-23 广东工业大学 Communication master user sensing method and system
KR20190119766A (en) * 2018-04-13 2019-10-23 서울과학기술대학교 산학협력단 Method for detecting position of radiation source using rotating modulation collimator and imaging device for performing the method
KR102042342B1 (en) 2018-04-13 2019-11-07 서울과학기술대학교 산학협력단 Method for detecting position of radiation source using rotating modulation collimator and imaging device for performing the method
CN115808648A (en) * 2022-11-18 2023-03-17 无锡鸣石峻致医疗科技有限公司 Device and method for measuring ringing noise of magnetic resonance system

Similar Documents

Publication Publication Date Title
CN109872288B (en) Network training method, device, terminal and storage medium for image denoising
US7978926B2 (en) Edge ringing artifact suppression methods and apparatuses
US6859564B2 (en) Signal processing using the self-deconvolving data reconstruction algorithm
JP2007188493A (en) Method and apparatus for reducing motion blur in motion blur image, and method and apparatus for generating image with reduced motion blur by using a plurality of motion blur images each having its own blur parameter
JP2001057677A (en) Image processing method, system and recording medium
JP5027757B2 (en) Moving image denoising device, method and program thereof
KR101707337B1 (en) Multiresolution non-local means filtering method for image denoising
JP2004362578A (en) Weighted overcomplete denoising
US10229479B2 (en) Image signal processing apparatus, image signal processing method and image signal processing program
US8055093B2 (en) Method and system for noise level detection in image data
Ponomaryov Real-time 2D–3D filtering using order statistics based algorithms
US8090215B2 (en) Method and system for noise level detection in image data
KR101464743B1 (en) Signal dependent noise estimation apparatus and method on camera module
JP2016083417A (en) Image processing device, image processing method, program, and computer recording medium
US20090257674A1 (en) Method and system for noise level detection in image data
Prakash et al. Medical image denoising based on soft thresholding using biorthogonal multiscale wavelet transform
Kaur et al. Image de-noising techniques: a review paper
CN110517196B (en) SAR image denoising method and system
Nitin A hybrid image denoising method based on discrete wavelet transformation with pre-gaussian filtering
Son et al. A pair of noisy/blurry patches-based PSF estimation and channel-dependent deblurring
Asha et al. Satellite image enhancement using contrast limited adaptive histogram equalization
Tuyet et al. Reducing impurities in medical images based on curvelet domain
Jain et al. Review on Denoising techniques for AWGN signal introduced in a stationary image
Chen et al. A structure-preserving image restoration method with high-level ensemble constraints
Lee et al. MMSE nonlocal means denoising algorithm for Poisson noise removal

Legal Events

Date Code Title Description
E701 Decision to grant or registration of patent right
GRNT Written decision to grant