CN108022220A - A kind of ultrasound pattern speckle noise minimizing technology - Google Patents
A kind of ultrasound pattern speckle noise minimizing technology Download PDFInfo
- Publication number
- CN108022220A CN108022220A CN201711275034.1A CN201711275034A CN108022220A CN 108022220 A CN108022220 A CN 108022220A CN 201711275034 A CN201711275034 A CN 201711275034A CN 108022220 A CN108022220 A CN 108022220A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msup
- mover
- image
- block
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
- 238000002604 ultrasonography Methods 0.000 title claims abstract description 21
- 238000005516 engineering process Methods 0.000 title claims abstract description 16
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 37
- 238000001914 filtration Methods 0.000 claims abstract description 22
- 238000002474 experimental method Methods 0.000 claims description 11
- 230000000694 effects Effects 0.000 claims description 10
- 238000009826 distribution Methods 0.000 claims description 6
- 230000007246 mechanism Effects 0.000 claims description 4
- 230000004927 fusion Effects 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000000034 method Methods 0.000 abstract description 12
- 238000012545 processing Methods 0.000 abstract description 9
- 230000008569 process Effects 0.000 abstract description 6
- 230000008901 benefit Effects 0.000 abstract description 5
- 238000003759 clinical diagnosis Methods 0.000 abstract description 3
- 230000009931 harmful effect Effects 0.000 abstract description 3
- 230000002401 inhibitory effect Effects 0.000 abstract 1
- 238000003384 imaging method Methods 0.000 description 6
- 238000012360 testing method Methods 0.000 description 6
- 238000011084 recovery Methods 0.000 description 4
- 239000000654 additive Substances 0.000 description 2
- 230000000996 additive effect Effects 0.000 description 2
- 210000003484 anatomy Anatomy 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 230000001605 fetal effect Effects 0.000 description 2
- 238000012804 iterative process Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 208000019155 Radiation injury Diseases 0.000 description 1
- 241000270295 Serpentes Species 0.000 description 1
- 210000001015 abdomen Anatomy 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013170 computed tomography imaging Methods 0.000 description 1
- 238000004195 computer-aided diagnosis Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001627 detrimental effect Effects 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 210000003754 fetus Anatomy 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 238000013173 literature analysis Methods 0.000 description 1
- 210000004185 liver Anatomy 0.000 description 1
- 210000005075 mammary gland Anatomy 0.000 description 1
- 210000000496 pancreas Anatomy 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 235000015170 shellfish Nutrition 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/10—Image enhancement or restoration using non-spatial domain filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The invention discloses a kind of ultrasound pattern speckle noise minimizing technology, and in iteration first, by the use of noise image as the input of model, a coarse despeckle image has been obtained after the filtering of bayesian non-local average filter model；In second of iteration, the input using acquired despeckle image as Filtering Model, obtains the more preferable despeckle image of a width；Iteration process, untill iterations reaches default value.Pass through the benign preferable despeckle image of one width of iterative filtering process final output.The advantage of the invention is that：The harmful effect for inhibiting speckle noise to be brought to clinical diagnosis or successive image processing.With reference to the statistical property of speckle noise, the conditional probability density function value in bayesian non-local average filter model is extrapolated, preferable despeckle image will be exported by benign iterative filtering process.In order to reduce the time complexity of algorithm, filtered using block, preselect block and control three kinds of means of iterations so that method can be practical.
Description
Technical field
The present invention relates to ultrasonoscopy processing technology field, more particularly to a kind of ultrasound pattern speckle noise minimizing technology.
Background technology
Ultrasonic imaging technique is a kind of conveniently imaging technique, is widely used among human body inspection, especially
It is the positions such as liver, courage, pancreas, abdomen, mammary gland.Compared with CT imaging techniques, ultrasonic imaging is much more secure, eliminates human body and is exposed to
Radiation injury caused by lower of X-ray is possible；Compared with MRI imaging techniques, the expense of ultrasonic imaging is much lower, be physical examination or
The important image mode of first run disorder in screening.However, the image acquired in ultrasonic instrument is all inevitably subject to spot at present
The interference of noise, this noise are caused by the coherence being imaged as ultrasonic system.The presence of speckle noise reduces figure
The resolution ratio and contrast of picture, reduce the quality of image, mask some detailed information, to clinical diagnosis and follow-up image
Processing analysis (such as：Feature extraction, lesion segmentation identification, image registration etc.) cause detrimental effect, therefore speckle noise
Suppression have very important significance to medical ultrasound image analysis.It is similar with other Image Denoising Technology problems, ultrasound figure
As the target of speckle noise reduction is should effectively to remove noise, retain the Key detail line in image as much as possible again
Manage feature.But speckle noise has certain difference with the Gaussian noise considered in usual image denoising problem, from the perspective of morphology
It appear that graininess or the effect of snake mark, its noise generation model is different from additive Gaussian from noise statistics
White noise, therefore directly existing denoising method can not be moved in the elimination of speckle noise, but need in denoising model
On fully take into account the characteristic of speckle noise, improve existing model, give full play to existing noise-removed technology on speckle noise reduction
Effect.
The content of the invention
The defects of present invention is directed to the prior art, there is provided a kind of ultrasound pattern speckle noise minimizing technology, can be effective
Solve the above-mentioned problems of the prior art.
In order to realize above goal of the invention, the technical solution that the present invention takes is as follows：
A kind of ultrasound pattern speckle noise minimizing technology, comprises the following steps that：
Step 1：
It is based on
In above formula, z is exactly the image containing speckle noise that B ultrasound instrument is obtained,It is the clean image of estimation, i is figure
The coordinate position of pixel in the Ω of image space；Δ (i) is represented centered on coordinate points i, size is the neighborhood search region of β × β, is sat
Punctuate j belongs in Δ (i)；Z (j) represent centered on pixel j in image z, size for α × α image block；Z (i) represent with
In image z centered on pixel i, the image block that size is α × α；P (z (i) | z (j)) in observation figure below of image block z (j)
The conditional probability density function value of picture block z (i),For in image blockObservation hypograph block z (i) bar
Part probability density function values；Represent estimating without spot for the image block centered on i；Estimated allAverage fusion is carried out, obtains despeckle image
Above formula implication is：In first step filtering, directly using noise image z as input, put down by the weighting of image block
Equal and combination can obtain the rough estimate to no spot image；Then, the input then using the estimation image filtered as next step,
Thus filter result is corrected；
According to formula 1 calculate p (z (i) | z (j)) under speckle noise environment andIt can be gone
Spot image, due toThe Gaussian noise that wherein n (i) is zero-mean, standard deviation is σ distributions, therefore
Z (i) | u (i)~N (u (i), u (i) σ2), then
Step 2：
In view of in image block pixel conditional sampling it is assumed that the probability distribution situation of image block can be obtained：
In above formula | R | represent image number of pixels in the block, r represents image r-th of pixel in the block.Based on formula 2, formula 1
P (z (i) | z (j)) in the first step may be calculated：
And in second stepIt can be calculated as：
Step 3：
In order to further lift despeckle performance, iterations is expanded to K times, (2≤K≤4), then complete iterative filtering
Despeckle model is shown below：
Preferably, the value of the h of adjustable type 5 is to obtain best despeckle effect in the step 3, due to h and σ2Correlation,
We make h=(C σ)2, summarized through many experiments, as C=1, despeckle effect is best.Can not be in the case of known σ, h takes
It is worth for 8 best results.
Preferably, the degree of overlapping between adjustment image block, reduces time complexity.
Preferably, block pre-selection mechanism is added in filtering, when the average and target image block that find the block in neighborhood are equal
When value difference is too big, then without considering contribution of the image block to object block estimate, i.e.,：
Default block average value threshold value is set to μ, makes μ=0.7；
WhenDuring beyond [μ, 1/ μ] scope, then willBlock pairContribution be set to 0.
Preferably, scope of the iterations K controls 2 to 4, and as speckle noise degree σ increases, K gradually drop
For 2, the time complexity that less iterations controls algorithm increases.
Compared with prior art the advantage of the invention is that：Largely inhibit speckle noise to clinical diagnosis or
The harmful effect that successive image processing is brought., will be defeated by benign iterative filtering process with reference to the statistical property of speckle noise
Go out preferable despeckle image.Meanwhile in order to reduce the time complexity of algorithm to greatest extent, filtered using block, preselect block and
Control three kinds of means of iterations so that method can be practical.
Brief description of the drawings
Fig. 1 is the middle despeckle that head noise phantom image (σ=10) and the method for the present invention produce in an iterative process
Figure；
Fig. 1 a are head noise phantom image (σ=10)；
Fig. 1 b are the result images after carried algorithm first time iteration
Fig. 1 c are the result images after carried second of iteration of algorithm
Fig. 1 d are the result images after carried algorithm third time iteration
Fig. 1 e are clean head phantom image；
Fig. 2 is despeckle result figure of the different despeckle algorithms to the head noise phantom image of σ=10；
Fig. 2 a are head noise phantom image (σ=10)；
Fig. 2 b are the despeckle result figure of SRAD；
Fig. 2 c are the despeckle result figure of SRBF；
Fig. 2 d are the despeckle result figure of TBNLMF；
Fig. 2 e are the despeckle result figure of OBNLMF；
Fig. 2 f put forward the despeckle result figure of algorithm for us；
Fig. 2 g are clean head phantom image；
Fig. 3 is emulation ultrasonic fetal image and despeckle image；
Fig. 3 a are the emulation ultrasonic fetal image of Field II Simulation Program generations；
Despeckle images of Fig. 3 b for carried algorithm to Fig. 3 a；
Fig. 4 is live ultrasound image and despeckle image；
Fig. 4 a are live ultrasound image；
The despeckle image that Fig. 4 b produce for carried algorithm.
Embodiment
For the objects, technical solutions and advantages of the present invention are more clearly understood, develop simultaneously embodiment referring to the drawings, right
The present invention is described in further details.
In traditional sense, speckle noise is considered as the multiplying property signal model of Rayleigh distributed, yet with image-forming mechanism
Complexity, this model be not suitable for true speckle noise model.It is more suitable in recent years, some scholars propose one kind
Speckle noise model, it is as follows：
Z (i)=u (i)+uγ(i)n(i). (1)
In above formula, z is exactly the image containing speckle noise that B ultrasound instrument is obtained, and u is it is desirable that what is recovered is clean
Image, the Gaussian noise that n is obedience zero-mean, standard deviation is σ distributions, i is the coordinate position of pixel in image space Ω, and γ is
With the relevant parameter of ultrasonic instrument, γ=0.5 is usually taken.From formula (1) as can be seen that speckle noise and additive Gaussian noise are not
With, it is impossible to algorithm directly is removed to handle speckle noise using Gaussian noise, it is necessary to fully takes into account the statistical property of noise
To change algorithm to adapt to speckle noise environment.
It is non local average filter under non-Gaussian noise environment according to a kind of bayesian non-local average filter model
Using providing possibility.The model is as follows：
Here, u (j) represent centered on pixel j in image u, size for α × α image block；Z (i) is represented with image
In z centered on pixel i, the image block that size is α × α；Δ (i) represent centered on coordinate points i, size for β × β neighborhood
Region of search, coordinate points j belong in Δ (i)；Represent estimating without spot for the image block centered on i.By all estimations
Go out(i ∈ Ω) carries out average fusion, obtains despeckle imageIn fact, this is a kind of noise based on image block
On filtering mode, with classical block-based non local average filter model form unanimously.
No spot image u is needed in apparent formula (2), and as inputting, this can not accomplish, therefore in order to realize the mould
Type, it is thus proposed that a kind of two step iterative models, by the process progressively to become more meticulous dexterously solve the problems, such as this：
The implication that above formula represents as：In first step filtering, directly using noise image z as input, it can obtain to nothing
The rough estimate of spot image；Then, the input then using the estimation image filtered as next step, thus corrects filter result.
Then, the speckle noise in ultrasonoscopy is removed using the model corresponding to formula (2), is made an uproar according to spot
The statistical property of sound is deduced the probability density function needed for model, obtains good despeckle effect.However, realizing shellfish
During the non local average filter models of Ye Si, existing classical documents eliminate follow-up the step of becoming more meticulous, and obtain despeckle image and compare
It is coarse.For this problem, this patent proposes a kind of ultrasound pattern speckle noise minimizing technology, specific as follows：
Based on model (3), we need to only calculate p (z (i) | z (j)) under speckle noise environment andIt can obtain despeckle image.Due toTherefore z (i) | u (i)~N (u (i),
u(i)σ2), then
In view of in image block pixel conditional sampling it is assumed that the probability distribution situation of image block can be obtained：
In above formula | R | represent image number of pixels in the block, r represents image r-th of pixel in the block.Based on (4), model
(3) p in the first step (z (i) | z (j)) may be calculated：
And in second stepIt can be calculated as：
In addition, in order to further lift despeckle performance, iterations is expanded to K times (2≤K≤4) by us, then completely
Iterative filtering despeckle model is as follows：
Notice that we instead of 2 σ with variable h in above formula2, in an experiment we adjust the value of h to obtain best despeckle
Effect.Final gainedAs exportDue to h and σ2Correlation, we make h=(C σ)2, summarized through many experiments,
As C=1, despeckle effect is best.Can not be in the case of known σ, h values be 8 best results.
In order to reduce the time complexity of Iterative-Filtering Scheme, this patent uses following three key processing mode：
(1) from formula (7) as can be seen that despeckle algorithm needs iteration K time, thus filter each time cannot spend it is too many
Time, otherwise the when consumption of algorithm is just long.In fact, the filter patterns based on image block can be by adjusting between image block
(the distance between adjacent image block center in other words, is set to d) consume during its filtering effectively to control degree of overlapping.According to pertinent literature
Analysis [12] understand：The time complexity of block-based non local average filter algorithm is O ((2 α+1)2(2β+1)2((N-d)/
d)2), N is the length of side of image here.If as it can be seen that the distance between image block d is arranged to 3, then it can make what is once filtered
Time shortens 9 times or so when than d being 1, and despeckle performance there will not be too big difference.Iteration K times, then can save 9K times
Time.
(2) block pre-selection mechanism is added in filtering, when the average for finding the block in neighborhood is differed with target image block average
, then can be without considering contribution of the image block to object block estimate, i.e., when too big：During beyond [μ, 1/ μ] scope
(μ is default block average value threshold value, usually makes μ=0.7), then willBlock pairContribution be set to 0, so not only shorten
It is corresponding to calculate the time, it also avoid harmful effect of the cumulative errors to finally estimating.
(3) scope of the iterations K controls 2 to 4, and as speckle noise degree σ increases, K are gradually reduced to 2, compared with
The time complexity that few iterations controls algorithm increases.
Experiment and interpretation of result：
In order to test the performance of put forward despeckle algorithm, we have chosen three kinds of test images, the first is according to being assumed
Noise model and the speckle noise image simulated, be for second analog simulation ultrasonoscopy, the third is straight from B ultrasound instrument
The real ultrasonoscopy obtained.Meanwhile have chosen several representational despeckle algorithms and contrasted with carried algorithm, they
It is respectively：SRAD,SRBF,traditional blockwise NLMF(TBNLMF),OBNLMF.For proposed by the invention
Iterative-Filtering Scheme, it is in order to obtain preferable despeckle effect, parameter setting is as follows：The range delta (i) of non local neighborhood is set
For 19 × 19；Image block is dimensioned to 5 × 5；As σ ＜ 10, iterations K is set to 4, the K=3 as 10≤σ ＜ 20, when
K=2 during σ >=20；H is arranged to σ in Filtering Model2；In addition, when σ is unknown, the value of h will be adjusted according to actual conditions, and
In view of conveniently factor, K is then set to 2.We verify the advantage of put forward algorithm by three groups of different experiments below.
Head noise phantom image experiment：
The head phantom image of several different noise levels is generated using the speckle noise modeling of hypothesis, for
These test images, due to the known corresponding clean image of immaculate, can use " full reference image quality appraisement index "
Quantitatively evaluating is done come the quality of the recovery image exported to algorithms of different.Here the index used is Y-PSNR (Peak
Signal-to-Noise Ratio, PSNR) and characteristic similarity Index (Feature Similarity Index
Measure, FSIM).The calculating formula of PSNR is：
Wherein MN is the size of image, FSIM is the image quality evaluation index developed in recent years, wherein image
Phase equalization and gradient amplitude are used for the difference for weighing image pair,.When recovery imageDuring closer to clean image u,
The value of FSIM is closer to 1.
First, in order to test the correctness of put forward iterative algorithm, we devise following experiment.In different noise levels
Under, caused recovery image in iterative process is all recorded, is denoted as respectivelyIt is right that its is calculated again
PSNR the and FSIM values answered, experimental result are as shown in table 1.Runic data in table are represented per the best restoration result value of a line.
There it can be seen that under any σ,Certainly ratioIt is better, and when σ is smaller, with the increase of iterations K,It can get over
It is better to come, this fully demonstrates the correctness of put forward iterative filtering despeckle algorithm.Meanwhile we illustrate propose calculation in Fig. 1
Produced during the head noise phantom image of method processing σ=10See each time with being apparent from
Makeover process of the iteration to recovery image.
Table 1 carries the middle denoising result (PSNRs, FSIMs) of iterative algorithm when handling " head noise phantom image "
Then, the head noise phantom image of different noise levels is input in selected despeckle algorithm by we, is obtained
Different despeckle images has been arrived, has calculated itself PSNR and FSIM value, the results are shown in Table 2.Equally, the runic data in table represent every
The best restoration result value of a line.As can be seen that carried algorithm is optimal under the picture quality deliberated index of PSNR and FSIM
Algorithm.Meanwhile we illustrate despeckle result figure of each algorithm to the head noise phantom image of σ=10 in fig. 2, from regarding
It can be seen that in feel：SRAD algorithms are stronger to the holding capacity of details, but many spots have still been retained in despeckle image；SRBF
Despeckle image in have the presence of many artifacts；TBNLMF algorithms can remove most spots, but lost many details
Feature；OBNLMF can keep reaching certain balance in blotch removal and details, but noise is still remained in image background
Trace；And the algorithm that we are proposed can not only remove most noises, while also done in the holding at details and edge
Obtain preferably.
The different despeckle algorithms of table 2 to the despeckle result of the head noise phantom image of different noise levels (PSNRs,
FSIMs)
The ultrasonoscopy experiment that Field II Simulation Program simulations produce：
Field II Simulation Program are that the ultrasonoscopy researched and developed according to ultrasonic imaging principle emulates generation
Program, has higher reference value.In the experiment of this section, we have chosen by Field II Simulation
" fetus " image of Program emulation generations, as our test image, as shown in Figure 3a.Then us are entered into
(pay attention in the despeckle algorithm proposed：Due to the information of no σ, 12) h in Filtering Model is set to, and the despeckle image of output is such as
Shown in Fig. 3 b.Two images are contrasted, we can become apparent from：Most speckle noises have been removed in right figure, together
When anatomical structure still complete display, this has absolutely proved the validity for carrying despeckle algorithm.
Live ultrasound image is tested：
The verification tested by first two, we have affirmed the validity for carrying algorithm, using true in the experiment of this section
Real ultrasonoscopy is as test object, as shown in fig. 4 a.Then recycle carry algorithm to the operation of carry out despeckle (attention：
In order to retain details as much as possible while despeckle, 8) h in Filtering Model is set to, gained image is as shown in Figure 4 b.From
Figure can be seen that：Despeckle image not remove only speckle noise, and the edge between adjacent anatomical structures is kept as very well,
This undoubtedly has very big benefit for the diagnosis of diagnostic imaging teacher.Certainly, to note here is that：Generally for medical image
For processing, we term it " computer-aided diagnosis ", doctor wishes that the image after combining artwork and handling is made jointly
Diagnostic result, rather than referring only to the figure after processing, because any type signal processing mode all can more or less lose
The potential minutia of signal, image contrast is with reference to so that diagnosis is more accurate.
Those of ordinary skill in the art will understand that the embodiments described herein, which is to help reader, understands this hair
Bright implementation, it should be understood that protection scope of the present invention is not limited to such special statement and embodiment.Ability
The those of ordinary skill in domain can according to the present invention these disclosed technical inspirations make it is various do not depart from essence of the invention its
Its various specific deformations and combination, these deformations and combination are still within the scope of the present invention.
Claims (5)
1. a kind of ultrasound pattern speckle noise minimizing technology, it is characterised in that comprise the following steps that：
Step 1：
It is based on
In above formula, z is exactly the image containing speckle noise that B ultrasound instrument is obtained,It is the clean image of estimation, i is image sky
Between in Ω pixel coordinate position；Δ (i) is represented centered on coordinate points i, size is the neighborhood search region of β × β, coordinate points
J belongs in Δ (i)；Z (j) represent centered on pixel j in image z, size for α × α image block；Z (i) is represented with image
In z centered on pixel i, the image block that size is α × α；P (z (i) | z (j)) in the observation hypograph block z of image block z (j)
(i) conditional probability density function value,For in image blockObservation hypograph block z (i) condition it is general
Rate density function values； Represent estimating without spot for the image block centered on i；Estimated allAverage fusion is carried out, obtains despeckle image
Above formula implication is：In first step filtering, directly using noise image z as inputting, by the weighted average of image block and
Combination can obtain the rough estimate to no spot image；Then, the input then using the estimation image filtered as next step, thus
Correct filter result；
According to formula 1 calculate p (z (i) | z (j)) under speckle noise environment andIt can obtain despeckle figure
Picture, due toThe Gaussian noise that wherein n (i) is zero-mean, standard deviation is σ distributions, therefore z (i)
| u (i)~N (u (i), u (i) σ2), then
Step 2：
In view of in image block pixel conditional sampling it is assumed that the probability distribution situation of image block can be obtained：
<mrow>
<mi>p</mi>
<mrow>
<mo>(</mo>
<mi>z</mi>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>|</mo>
<mi>u</mi>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mo>&Proportional;</mo>
<msubsup>
<mo>&Pi;</mo>
<mrow>
<mi>r</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mo>|</mo>
<mi>R</mi>
<mo>|</mo>
</mrow>
</msubsup>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<mfrac>
<msup>
<mrow>
<mo>(</mo>
<msup>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>-</mo>
<msup>
<mi>u</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mrow>
<mn>2</mn>
<msup>
<mi>u</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<msup>
<mi>&sigma;</mi>
<mn>2</mn>
</msup>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>r</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mo>|</mo>
<mi>R</mi>
<mo>|</mo>
</mrow>
</msubsup>
<mfrac>
<msup>
<mrow>
<mo>(</mo>
<msup>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>-</mo>
<msup>
<mi>u</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mrow>
<mn>2</mn>
<msup>
<mi>u</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<msup>
<mi>&sigma;</mi>
<mn>2</mn>
</msup>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</mrow>
In above formula | R | represent image number of pixels in the block, r represents image r-th of pixel in the block, based on formula 2, formula 1 first
P (z (i) | z (j)) in step may be calculated：
<mrow>
<mi>p</mi>
<mrow>
<mo>(</mo>
<mi>z</mi>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>|</mo>
<mi>z</mi>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>r</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mo>|</mo>
<mi>R</mi>
<mo>|</mo>
</mrow>
</msubsup>
<mfrac>
<msup>
<mrow>
<mo>(</mo>
<msup>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>-</mo>
<msup>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mrow>
<mn>2</mn>
<msup>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<msup>
<mi>&sigma;</mi>
<mn>2</mn>
</msup>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mo>.</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
</mrow>
And in second stepIt can be calculated as：
<mrow>
<mi>p</mi>
<mrow>
<mo>(</mo>
<mi>z</mi>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>|</mo>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>r</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mo>|</mo>
<mi>R</mi>
<mo>|</mo>
</mrow>
</msubsup>
<mfrac>
<msup>
<mrow>
<mo>(</mo>
<msup>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>-</mo>
<msup>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mrow>
<mn>2</mn>
<msup>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<msup>
<mi>&sigma;</mi>
<mn>2</mn>
</msup>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mo>.</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mrow>
Step 3：
In order to further lift despeckle performance, iterations is expanded to K times, (2≤K≤4), then complete iterative filtering despeckle
Model is shown below：
<mrow>
<mtable>
<mtr>
<mtd>
<mrow>
<mi>I</mi>
<mi>n</mi>
<mi>i</mi>
<mi>t</mi>
<mi>i</mi>
<mi>a</mi>
<mi>l</mi>
<mi>i</mi>
<mi>z</mi>
<mi>e</mi>
<mo>:</mo>
<mi>k</mi>
<mo>=</mo>
<mn>1.</mn>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>W</mi>
<mi>h</mi>
<mi>i</mi>
<mi>l</mi>
<mi>e</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>&le;</mo>
<mi>K</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>i</mi>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>=</mo>
<mo>=</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>:</mo>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>&Element;</mo>
<mi>&Delta;</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mo>&CenterDot;</mo>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>r</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mo>|</mo>
<mi>R</mi>
<mo>|</mo>
</mrow>
</msubsup>
<msup>
<mrow>
<mo>(</mo>
<mrow>
<msup>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<msup>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>/</mo>
<msup>
<mi>hz</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msub>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>&Element;</mo>
<mi>&Delta;</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>r</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mo>|</mo>
<mi>R</mi>
<mo>|</mo>
</mrow>
</msubsup>
<msup>
<mrow>
<mo>(</mo>
<mrow>
<msup>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<msup>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>/</mo>
<msup>
<mi>hz</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&DoubleRightArrow;</mo>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mo>;</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>k</mi>
<mo>=</mo>
<mi>k</mi>
<mo>+</mo>
<mn>1.</mn>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>e</mi>
<mi>l</mi>
<mi>s</mi>
<mi>e</mi>
<mo>:</mo>
<mover>
<mi>u</mi>
<mo>~</mo>
</mover>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>&Element;</mo>
<mi>&Delta;</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mrow>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mo>&CenterDot;</mo>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>r</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mo>|</mo>
<mi>R</mi>
<mo>|</mo>
</mrow>
</msubsup>
<msup>
<mrow>
<mo>(</mo>
<mrow>
<msup>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<msup>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>/</mo>
<mi>h</mi>
<msup>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msub>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>&Element;</mo>
<mi>&Delta;</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>r</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mo>|</mo>
<mi>R</mi>
<mo>|</mo>
</mrow>
</msubsup>
<msup>
<mrow>
<mo>(</mo>
<mrow>
<msup>
<mi>z</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<msup>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>/</mo>
<mi>h</mi>
<msup>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&DoubleRightArrow;</mo>
<mover>
<mi>u</mi>
<mo>~</mo>
</mover>
<mo>;</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>k</mi>
<mo>=</mo>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
<mo>;</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mo>=</mo>
<mover>
<mi>u</mi>
<mo>~</mo>
</mover>
<mo>.</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mrow>
<mi>o</mi>
<mi>p</mi>
<mi>t</mi>
</mrow>
</msub>
<mo>=</mo>
<mover>
<mi>u</mi>
<mo>^</mo>
</mover>
<mo>.</mo>
</mrow>
</mtd>
</mtr>
</mtable>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
<mo>.</mo>
</mrow>
A kind of 2. ultrasound pattern speckle noise minimizing technology according to claim 1, it is characterised in that：In the step 3
The value of the h of adjustable type 5 is to obtain best despeckle effect, due to h and σ2Correlation, makes h=(C σ)2, summarized through many experiments,
As C=1, despeckle effect is best, and can not be in the case of known σ, h values be 8 best results.
A kind of 3. ultrasound pattern speckle noise minimizing technology according to claim 1, it is characterised in that：Adjust image block it
Between degree of overlapping, reduce time complexity.
A kind of 4. ultrasound pattern speckle noise minimizing technology according to claim 1, it is characterised in that：Added in filtering
Block preselects mechanism, when the average for finding the block in neighborhood differs too big with target image block average, then without considering the image block
Contribution to object block estimate, i.e.,：
Default block average value threshold value is set to μ, makes μ=0.7；
WhenDuring beyond [μ, 1/ μ] scope, then willBlock pairContribution be set to 0.
A kind of 5. ultrasound pattern speckle noise minimizing technology according to claim 1, it is characterised in that：Iterations K is controlled
The scope 2 to 4 is made, and as speckle noise degree σ increases, K are gradually reduced to 2, less iterations controls algorithm
Time complexity increase.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711275034.1A CN108022220B (en) | 2017-12-06 | 2017-12-06 | Ultrasonic image speckle noise removing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711275034.1A CN108022220B (en) | 2017-12-06 | 2017-12-06 | Ultrasonic image speckle noise removing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108022220A true CN108022220A (en) | 2018-05-11 |
| CN108022220B CN108022220B (en) | 2021-04-20 |
Family
ID=62078607
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711275034.1A Active CN108022220B (en) | 2017-12-06 | 2017-12-06 | Ultrasonic image speckle noise removing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108022220B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109087266A (en) * | 2018-08-09 | 2018-12-25 | 苏州大学 | A kind of image speckle iteration reduction method of combination bilateral filtering |
| CN110390646A (en) * | 2019-06-12 | 2019-10-29 | 西南科技大学 | A kind of details holding image de-noising method |
| CN114037616A (en) * | 2021-09-22 | 2022-02-11 | 南京莱斯电子设备有限公司 | SAR image noise suppression method and equipment |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5594807A (en) * | 1994-12-22 | 1997-01-14 | Siemens Medical Systems, Inc. | System and method for adaptive filtering of images based on similarity between histograms |
| CN101661611A (en) * | 2009-09-25 | 2010-03-03 | 西安电子科技大学 | Realization method based on bayesian non-local mean filter |
| US7744534B2 (en) * | 2002-06-07 | 2010-06-29 | Verathon Inc. | 3D ultrasound-based instrument for non-invasive measurement of amniotic fluid volume |
| CN102609904A (en) * | 2012-01-11 | 2012-07-25 | 云南电力试验研究院（集团）有限公司电力研究院 | Bivariate nonlocal average filtering de-noising method for X-ray image |
| CN103886563A (en) * | 2014-04-18 | 2014-06-25 | 中南民族大学 | SAR image speckle noise inhibition method based on non-local mean and heterogeneity measurement |
| US20140193095A1 (en) * | 2013-01-04 | 2014-07-10 | Samsung Electronics Co., Ltd. | Method and apparatus for image correction |
| CN103927715A (en) * | 2014-03-14 | 2014-07-16 | 中瑞科技（常州）有限公司 | Ultrasound image speckle noise suppression method |
| US8928813B2 (en) * | 2010-10-28 | 2015-01-06 | Microsoft Corporation | Methods and apparatus for reducing structured noise in video |
| CN107220941A (en) * | 2016-12-15 | 2017-09-29 | 西南科技大学 | A kind of method and system for removing image impact noise |
-
2017
- 2017-12-06 CN CN201711275034.1A patent/CN108022220B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5594807A (en) * | 1994-12-22 | 1997-01-14 | Siemens Medical Systems, Inc. | System and method for adaptive filtering of images based on similarity between histograms |
| US7744534B2 (en) * | 2002-06-07 | 2010-06-29 | Verathon Inc. | 3D ultrasound-based instrument for non-invasive measurement of amniotic fluid volume |
| CN101661611A (en) * | 2009-09-25 | 2010-03-03 | 西安电子科技大学 | Realization method based on bayesian non-local mean filter |
| US8928813B2 (en) * | 2010-10-28 | 2015-01-06 | Microsoft Corporation | Methods and apparatus for reducing structured noise in video |
| CN102609904A (en) * | 2012-01-11 | 2012-07-25 | 云南电力试验研究院（集团）有限公司电力研究院 | Bivariate nonlocal average filtering de-noising method for X-ray image |
| US20140193095A1 (en) * | 2013-01-04 | 2014-07-10 | Samsung Electronics Co., Ltd. | Method and apparatus for image correction |
| CN103927715A (en) * | 2014-03-14 | 2014-07-16 | 中瑞科技（常州）有限公司 | Ultrasound image speckle noise suppression method |
| CN103886563A (en) * | 2014-04-18 | 2014-06-25 | 中南民族大学 | SAR image speckle noise inhibition method based on non-local mean and heterogeneity measurement |
| CN107220941A (en) * | 2016-12-15 | 2017-09-29 | 西南科技大学 | A kind of method and system for removing image impact noise |
Non-Patent Citations (3)
| Title |
|---|
| PIERRICK COUP´等: "BAYESIAN NON LOCAL MEANS-BASED SPECKLE FILTERING", 《IEEE INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING: FROM NANO TO MACRO》 * |
| YUMING FANG等: "Video Saliency Incorporating Spatiotemporal Cues and Uncertainty Weighting", 《IEEE TRANSACTIONS ON IMAGE PROCESSING》 * |
| 韩寒等: "超声图像去噪方法", 《计算机工程与应用》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109087266A (en) * | 2018-08-09 | 2018-12-25 | 苏州大学 | A kind of image speckle iteration reduction method of combination bilateral filtering |
| CN110390646A (en) * | 2019-06-12 | 2019-10-29 | 西南科技大学 | A kind of details holding image de-noising method |
| CN110390646B (en) * | 2019-06-12 | 2023-10-17 | 西南科技大学 | Detail-preserving image denoising method |
| CN114037616A (en) * | 2021-09-22 | 2022-02-11 | 南京莱斯电子设备有限公司 | SAR image noise suppression method and equipment |
| CN114037616B (en) * | 2021-09-22 | 2024-04-05 | 南京莱斯电子设备有限公司 | SAR image noise suppression method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108022220B (en) | 2021-04-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11478212B2 (en) | Method for controlling scanner by estimating patient internal anatomical structures from surface data using body-surface and organ-surface latent variables | |
| Singh et al. | Feature enhancement in medical ultrasound videos using contrast-limited adaptive histogram equalization | |
| US8842936B2 (en) | Method, apparatus, and program for aligning images | |
| KR101017611B1 (en) | System and method for extracting anatomical feature | |
| De Craene et al. | 3D strain assessment in ultrasound (straus): A synthetic comparison of five tracking methodologies | |
| US8447383B2 (en) | System and method for providing 2-dimensional computerized-tomography image corresponding to 2-dimensional ultrasound image | |
| CN108022220A (en) | A kind of ultrasound pattern speckle noise minimizing technology | |
| US8494239B2 (en) | Image processing device, method and program | |
| CN105813572A (en) | Image compounding based on image information | |
| Sakalauskas et al. | Automated segmentation of transcranial sonographic images in the diagnostics of Parkinson’s disease | |
| Ahn et al. | Adaptive template filtering for signal-to-noise ratio enhancement in magnetic resonance imaging | |
| CN109087266A (en) | A kind of image speckle iteration reduction method of combination bilateral filtering | |
| Sahli et al. | Analytic approach for fetal head biometric measurements based on log gabor features | |
| Sun et al. | Subpixel registration in renal perfusion MR image sequence | |
| CN111815527A (en) | Mixed high-order variational ultrasonic image denoising method based on Weibull distribution | |
| Agrawal et al. | Development of adaptive fuzzy based Image Filtering techniques for efficient Noise Reduction in Medical Images | |
| JP4571378B2 (en) | Image processing method, apparatus, and program | |
| Kumar et al. | Semiautomatic method for segmenting pedicles in vertebral radiographs | |
| Pan et al. | Carpal-bone feature extraction analysis in skeletal age assessment based on deformable model | |
| CN111242853A (en) | Medical CT image denoising method based on optical flow processing | |
| Yousefi et al. | A fully automated segmentation of radius bone based on active contour in wrist MRI data set | |
| Gupta et al. | Comparative Analysis of Anisotropic Diffusion and Non-Local Means on Ultrasound Images | |
| Sriraam et al. | Performance evaluation of linear and nonlinear filters for despeckling B mode foetal heart ultrasound images | |
| Chumchob et al. | An improved variational model and its numerical solutions for speckle noise removal from real ultrasound images | |
| Lee et al. | Geometric active model for lesion segmentation on breast ultrasound images |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |