CN107368851A - A kind of Fast Fuzzy C-Means Clustering image partition method with neighborhood choice strategy - Google Patents
A kind of Fast Fuzzy C-Means Clustering image partition method with neighborhood choice strategy Download PDFInfo
- Publication number
- CN107368851A CN107368851A CN201710558682.1A CN201710558682A CN107368851A CN 107368851 A CN107368851 A CN 107368851A CN 201710558682 A CN201710558682 A CN 201710558682A CN 107368851 A CN107368851 A CN 107368851A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- window
- neighborhood
- pixel
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
Abstract
The invention discloses a kind of Fast Fuzzy C-Means Clustering image partition method with neighborhood choice strategy; including carrying out edge extracting to image; the window and its neighborhood of suitable size are selected using the marginal information of extraction; window neighborhood choice strategy is introduced in local similarity and new figure ξ calculating process; and suitable cluster centre is obtained using particle cluster algorithm; improve traditional fuzzy C means clustering methods so that image segmentation details is preferably protected.
Description
Technology neighborhood
The present invention relates to image to split neighborhood, and in particular to a kind of Fast Fuzzy C-Means Clustering with neighborhood choice strategy
Image partition method.
Background technology
The details protection of image is all one of emphasis and difficult point of image segmentation all the time, and traditional Fast Fuzzy C is equal
It is not fine that value cluster image partition method (FGFCM), which retains edge effect, does not account for window neighborhood choice and edge
The relation of information, only select neighborhood point corresponding to the window and its object pixel of fixed size.
The content of the invention
In order to overcome shortcoming and deficiency existing for prior art, the present invention provides a kind of Fast Modular with neighborhood choice strategy
Paste C mean cluster image partition methods.
It is of the invention mainly using to image progress edge extracting, the marginal information of extraction come select the window of suitable size and
Its neighborhood, window neighborhood choice strategy is introduced in local similarity and new figure ξ calculating process, and uses particle cluster algorithm
(PSO) suitable cluster centre is obtained, improves traditional fuzzy C means clustering methods so that image segmentation details obtains more preferably
Protection.
The present invention adopts the following technical scheme that:
A kind of Fast Fuzzy C-Means Clustering image partition method with neighborhood choice strategy, comprises the following steps:
S1 inputs picture to be split, extracts image border;
Neighborhood choice strategy generatings of the S2 based on image border newly schemes ξ;
S3 obtains initial cluster center according to the new figure of generation;
S4 Fast Fuzzy C-Means Clusterings obtain optimal clustering.
Neighborhood choice strategy generatings of the S2 based on image border newly schemes ξ, is specially:
S2.1 is based on image border and carries out window neighborhood choice;
S2.2 combinations space and gray-scale information obtain local similarity amount, and calculation formula is as follows:
I pixels are the centers of local window, and k pixels represent the pixel in the window neighborhood of i pixels, and the window neighborhood is base
Obtained in the window neighborhood choice strategy described in S2.1, pi, qiIt is pixel i coordinate, xiIt is the gray value of window neighborhood, λs
And λgIt is two scale factors;
σiIt is defined as:
S2.3 calculates the new figure ξ of generation
ξ is calculated as follows shown in formula
Wherein, ξiRepresent the gray value of figure ξ ith pixel, xkRepresent x in artworkiThe gray value of neighborhood territory pixel, the window
Mouth neighborhood is obtained based on the window neighborhood choice strategy described in S2.1, NiIt is xiNeighborhood collection, SikIt is ith pixel and
Local similarity amount between k pixel.
S2.1 is based on image border and carries out window neighborhood choice, is specially:
Initial window size is set as 5*5, if non-flanged is fallen within the window, selects the window as local window,
Pixel in window is target pixel neighborhood;
If edge be present in the window, window is expanded as into 7*7, selection and the pixel of object pixel edge homonymy are made
For neighborhood.
The λsAnd λgIt is both configured to 2.
Initial cluster center parameter is obtained using particle cluster algorithm.
The S4 Fast Fuzzy C-Means Clusterings obtain optimal clustering, J values is reached minimum and obtain optimal cluster
Division
Wherein, cjIt is jth class center, uijIt is that the pixel that gray value is i belongs to the degree of membership of jth class, M is figure ξ ash
Spend series, γiIt is the pixel number with i with gray value, m is the Fuzzy Exponential factor, ξiRepresent the gray scale of figure ξ ith pixel
Value;
uijWith cjBetween iterative relation be shown below:
Detailed process is:
S4.1 is initialized, and sets Fuzzy Exponential factor m, primary iteration to count b2, highest iterations t2, iteration threshold ε
Parameter, cluster subordinated-degree matrix initial value U is set(0)=Ubest, initial cluster center
S4.2 utilizes cjUpdate class center cj, utilize uijUpdate subordinated-degree matrix U (b+1);
If S4.3 max { U (b)-U (b+1) }<ε or b2 > t2, then iteration stopping, otherwise, b=b+1, continues step S4.2.
Fuzzy Exponential factor m is set to 2, and iteration count initialization b2=0, it is 100 times to set highest iterations t2, iteration
Terminate threshold epsilon and be set to 1e-5.
Beneficial effects of the present invention:
The present invention adds the window based on edge in the method that traditional Fast Fuzzy C-Means Clustering image is split
Neighborhood choice strategy, the details protection to image segmentation serve the effect of enhancing.
Brief description of the drawings
Fig. 1 is the flow chart that the present invention works;
Fig. 2 is searching class central flow figure of the present invention;
Fig. 3 is window neighborhood choice strategic process figure;
Fig. 4 is cluster flow chart;
Fig. 5 (a) is artwork, and Fig. 5 (b) is standard drawing, and Fig. 5 (c) is using conventional method FGFCM effect picture, Fig. 5
(d) picture to be split using this method.
Embodiment
With reference to embodiment and accompanying drawing, the present invention is described in further detail, but embodiments of the present invention are not
It is limited to this.
Embodiment
As Figure 1-Figure 4, a kind of Fast Fuzzy C-Means Clustering image partition method with neighborhood choice strategy, including
Following steps:
S1 inputs picture to be split, extracts image border;
Because Canny edge detecting operations are more convenient and effect is preferable, Canny rim detections are selected, extract image border.
Neighborhood choice strategy generatings of the S2 based on image border newly schemes ξ, is specially:
S2.1 selects neighborhood by the relation at window and edge, can strengthen the protection to image detail, the choosing of window neighborhood
Select, specific strategy is as follows:
(1) initial window size is set as 5*5, if non-flanged is fallen within the window, selects the window as local window
Mouthful, the pixel in window is target pixel neighborhood;
(2) set initial window size as 5*5, if edge be present in the window, window expanded as into 7*7, selection with
The pixel of object pixel edge homonymy is as neighborhood.
S2.2 calculates local similarity;
Obtain local similarity amount with reference to space and gray-scale information, calculation formula is as follows, set ratio in formula because
Sub- λs、λg;
I pixels are the centers of local window, and k pixels represent the pixel in the window neighborhood of i pixels, and the window neighborhood is base
Obtained in the window neighborhood choice strategy described in S2.1, pi, qiIt is pixel i coordinate, xiIt is the gray value of window neighborhood, λs
And λgIt is two scale factors;
σiIt is defined as:
N in formula (2)iFor the window neighborhood territory pixel set of i pixels, NRFor the intraoral neighborhood territory pixel total number of i pixel windows, the window
Mouth neighborhood is obtained based on the window neighborhood choice strategy described in step 1.
In the embodiment of the present invention, λsAnd λgIt is both configured to 2.
S2.3 calculates the new figure ξ of generation
ξ is calculated as follows shown in formula
Wherein, ξiRepresent the gray value of figure ξ ith pixel.xkRepresent x in artworkiThe gray value of neighborhood territory pixel, the window
Mouth neighborhood is obtained based on the window neighborhood choice strategy described in step 1, NiIt is xiNeighborhood collection.SikBe ith pixel and
Local similarity amount between k-th of pixel.
S2.4 obtains initial cluster center according to the new figure of generation;
Select different cluster centres to split to image and play different effects, and in order to avoid being absorbed in local optimum,
The present invention obtains preferable initial cluster center parameter using particle cluster algorithm, and particle cluster algorithm is according to the more new particle of formula 7,8
Position and speed.
vij=wvij+c1·rand1ij·(pbestij-xij)+c2·rand2ij·(gbestj-xij) (7)
xij=xij+vij (8)
Wherein i=1,2 ..., M, M are the sums of particle in the colony;VijIt is speed of i-th of particle in jth time circulation
Degree;xijFor position of i-th of particle in jth time circulation, pbesijAnd gbestjRespectively i-th of particle is in j circulation
Optimum position and global optimum's particle position of j circulation;Rand1ij and rand2ij is the random number between (0,1);
C1 and c2 is Studying factors, and w is inertial factor.
The judge criterion of particle cluster algorithm represents that the fitness function used in the present invention is arranged to fitness function
Shown in following formula 9, when fitness value is bigger, then illustrate that particle position is more excellent.
Fitness=1/ (1+J) (9)
As shown in Fig. 2 initial cluster center search routine, specific search procedure is as follows:
S2.4.1 parameter settings
Set population quantity, population quantity, Studying factors c1 and c2, highest iterations t1, iteration count initialization b1,
Cluster number c.
In simulation example Fig. 5 of the present invention, population quantity is arranged to 50, and population quantity is arranged to 0.5, Studying factors c1 and
C2 is both configured to 0.5, and highest iterations t1 is 1000 times, and iteration count initialization b1=1, cluster number c are set to 3.
S2.4.2 is randomly provided particle original position
The each particles of S2.4.3 calculate degree of membership by formula 5, and fitness is calculated by formula 9, find each particle and follow for b1 times
In ring optimal location and b1 times circulation in global optimum's particle position, utilize formula 10 and 11 update particle position.
If S2.4.4 iterations b1 < t1, iterations b1=b1+1 is set, continues step (3);As b1 >=t1, i.e.,
When iterations reaches the limitation of highest iterations, stop iteration, by global optimum's particle of acquisition
Initial cluster center of the position as Fast Fuzzy C-Means Clustering
And obtain its subordinated-degree matrix U by 5 formulasbest。
S4 Fast Fuzzy C-Means Clusterings obtain optimal clustering;
On the basis of new figure, Fast Fuzzy C-Means Clustering is by minimizing formula 4, to obtain optimal clustering.
Wherein, cj is jth class center, and uij is that the pixel that gray value is i belongs to the degree of membership of jth class, and M is figure ξ ash
Spend series, γiIt is the pixel number with i with gray value, m is the Fuzzy Exponential factor, iterative relation such as following formula 5,6 between uij and cj
It is shown:
Shown in cluster comprises the following steps that:
(1) initialize
Formula (4), (5), the Fuzzy Exponential factor m in (6), primary iteration is set to count b2, highest iterations t2, repeatedly
For threshold epsilon parameter.On the basis of step 3 obtains new figure, the U obtained in step 4 is utilizedbestAnd cvest, cluster degree of membership is set
Matrix setup values U(0)=Ubest, initial cluster center
In the present invention, Fuzzy Exponential factor m is set to 2.Iteration count initializes b2=0, and setting highest iterations t2 is
100 times, iteration ends threshold epsilon is set to 1e-5.
(2) class center c is updated according to formula 6j, update subordinated-degree matrix U (b+1) using formula 5.
(3) if max { U (b)-U (b+1) }<ε or b2 > t2, then iteration stopping, otherwise, b=b+1, continues step (2)
The standard that image segmentation judges, that is, split accuracy rate (SA) calculation formula, contrasted with conventional method, illustrate this
The accuracy rate of method is higher, and concrete outcome is shown in Table 1, and the table is the average results of 10 experimental calculations.
In formula, C is cluster kind of a number, and Aj is the pixel point set for belonging to J classes that is found by certain dividing method, and Cj is
Belong to the pixel point set of J classes in standard drawing, if SA value is higher, illustrate that segmentation effect is better.Ideally, should
It is worth for 1.
Table 1
FGFCM | Context of methods |
0.9770 | 0.9820 |
Carried out using the inventive method shown in image segmentation matlab simulated effect figures, Fig. 5 (a) artworks, Fig. 5 (b) is mark
Quasi- figure, Fig. 5 (c) are the effect picture using this conventional method FGFCM, and Fig. 5 (d) is the picture split using this method, can be sent out
Existing the inventive method image segmentation is preferable.
Above-described embodiment is the preferable embodiment of the present invention, but embodiments of the present invention are not by the embodiment
Limitation, other any Spirit Essences without departing from the present invention with made under principle change, modification, replacement, combine, simplification,
Equivalent substitute mode is should be, is included within protection scope of the present invention.
Claims (8)
1. a kind of Fast Fuzzy C-Means Clustering image partition method with neighborhood choice strategy, it is characterised in that including following step
Suddenly:
S1 inputs picture to be split, extracts image border;
Neighborhood choice strategy generatings of the S2 based on image border newly schemes ξ;
S3 obtains initial cluster center according to the new figure of generation;
S4 Fast Fuzzy C-Means Clusterings obtain optimal clustering.
2. Fast Fuzzy C-Means Clustering image partition method according to claim 1, it is characterised in that the S2 is based on
The neighborhood choice strategy generating of image border newly schemes ξ, is specially:
S2.1 is based on image border and carries out window neighborhood choice;
S2.2 combinations space and gray-scale information obtain local similarity amount, and calculation formula is as follows:
<mrow>
<msub>
<mi>s</mi>
<mrow>
<mi>i</mi>
<mi>k</mi>
</mrow>
</msub>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mi>max</mi>
<mrow>
<mo>(</mo>
<mo>|</mo>
<msub>
<mi>p</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>p</mi>
<mi>k</mi>
</msub>
<mo>|</mo>
<mo>,</mo>
<mo>|</mo>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>q</mi>
<mi>k</mi>
</msub>
<mo>|</mo>
<mo>)</mo>
</mrow>
<mo>/</mo>
<msub>
<mi>&lambda;</mi>
<mi>s</mi>
</msub>
<mo>-</mo>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>x</mi>
<mi>k</mi>
</msub>
<mo>|</mo>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>/</mo>
<msub>
<mi>&lambda;</mi>
<mi>g</mi>
</msub>
<msubsup>
<mi>&sigma;</mi>
<mi>i</mi>
<mn>2</mn>
</msubsup>
</mrow>
</msup>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mo>&NotEqual;</mo>
<mi>k</mi>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mn>0</mn>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mi>k</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
I pixels are the centers of local window, and k pixels represent the pixel in the window neighborhood of i pixels, and the window neighborhood is to be based on
What the window neighborhood choice strategy described in S2.1 obtained, pi, qiIt is pixel i coordinate, xiIt is the gray value of window neighborhood, λsAnd λg
It is two scale factors;
σiIt is defined as:
<mrow>
<msub>
<mi>&sigma;</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<msqrt>
<mfrac>
<mrow>
<munder>
<mo>&Sigma;</mo>
<mrow>
<mi>k</mi>
<mo>&Element;</mo>
<msub>
<mi>N</mi>
<mi>i</mi>
</msub>
</mrow>
</munder>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>x</mi>
<mi>k</mi>
</msub>
<mo>|</mo>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
<msub>
<mi>N</mi>
<mi>R</mi>
</msub>
</mfrac>
</msqrt>
</mrow>
S2.3 calculates the new figure ξ of generation
ξ is calculated as follows shown in formula
<mrow>
<msub>
<mi>&xi;</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<munder>
<mo>&Sigma;</mo>
<mrow>
<mi>k</mi>
<mo>&Element;</mo>
<msub>
<mi>N</mi>
<mi>i</mi>
</msub>
</mrow>
</munder>
<msub>
<mi>S</mi>
<mrow>
<mi>i</mi>
<mi>k</mi>
</mrow>
</msub>
<msub>
<mi>x</mi>
<mi>k</mi>
</msub>
</mrow>
<mrow>
<munder>
<mo>&Sigma;</mo>
<mrow>
<mi>k</mi>
<mo>&Element;</mo>
<msub>
<mi>N</mi>
<mi>i</mi>
</msub>
</mrow>
</munder>
<msub>
<mi>S</mi>
<mrow>
<mi>i</mi>
<mi>k</mi>
</mrow>
</msub>
</mrow>
</mfrac>
</mrow>
Wherein, ξiRepresent the gray value of figure ξ ith pixel, xkRepresent x in artworkiThe gray value of neighborhood territory pixel, the window are adjacent
Domain is obtained based on the window neighborhood choice strategy described in S2.1, NiIt is xiNeighborhood collection, SikIt is ith pixel and k-th
Local similarity amount between pixel.
3. Fast Fuzzy C-Means Clustering image partition method according to claim 1, it is characterised in that S2.1 is based on figure
As edge progress window neighborhood choice, it is specially:
Initial window size is set as 5*5, if non-flanged is fallen within the window, selects the window as local window, window
Interior pixel is target pixel neighborhood;
If edge be present in the window, window is expanded as into 7*7, selection and the pixel of object pixel edge homonymy are used as neighbour
Domain.
4. Fast Fuzzy C-Means Clustering image partition method according to claim 2, it is characterised in that the λsAnd λgAll
It is arranged to 2.
5. Fast Fuzzy C-Means Clustering image partition method according to claim 1, it is characterised in that using population
Algorithm obtains initial cluster center parameter.
6. Fast Fuzzy C-Means Clustering image partition method according to claim 1, it is characterised in that the S4 is quick
Fuzzy C-means clustering obtains optimal clustering, J values is reached minimum and obtains optimal clustering
<mrow>
<mi>J</mi>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>M</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>c</mi>
</munderover>
<msub>
<mi>&gamma;</mi>
<mi>i</mi>
</msub>
<msubsup>
<mi>u</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mi>m</mi>
</msubsup>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>&xi;</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>c</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
Wherein, cjIt is jth class center, uijIt is that the pixel that gray value is i belongs to the degree of membership of jth class, M is figure ξ gray level
Number, γiIt is the pixel number with i with gray value, m is the Fuzzy Exponential factor, ξiRepresent the gray value of figure ξ ith pixel;
uijWith cjBetween iterative relation be shown below:
<mrow>
<msub>
<mi>u</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>&xi;</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>c</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mfrac>
<mn>2</mn>
<mrow>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</mfrac>
</msup>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>k</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>c</mi>
</munderover>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>&xi;</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>c</mi>
<mi>k</mi>
</msub>
<mo>)</mo>
</mrow>
<mfrac>
<mn>2</mn>
<mrow>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</mfrac>
</msup>
</mrow>
</mfrac>
</mrow>
<mrow>
<msub>
<mi>c</mi>
<mi>j</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>M</mi>
</munderover>
<msub>
<mi>&gamma;</mi>
<mi>i</mi>
</msub>
<msubsup>
<mi>u</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mi>m</mi>
</msubsup>
<msub>
<mi>&xi;</mi>
<mi>i</mi>
</msub>
</mrow>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>M</mi>
</munderover>
<msub>
<mi>&gamma;</mi>
<mi>i</mi>
</msub>
<msubsup>
<mi>u</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mi>m</mi>
</msubsup>
</mrow>
</mfrac>
<mo>.</mo>
</mrow>
7. Fast Fuzzy C-Means Clustering image partition method according to claim 6, it is characterised in that detailed process is:
S4.1 is initialized, and sets Fuzzy Exponential factor m, primary iteration to count b2, highest iterations t2, iteration threshold ε parameters,
Cluster subordinated-degree matrix initial value U is set(0)=Ubest, initial cluster center
S4.2 utilizes cjUpdate class center cj, utilize uijUpdate subordinated-degree matrix U (b+1);
If S4.3 max { U (b)-U (b+1) }<ε or b2 > t2, then iteration stopping, otherwise, b=b+1, continues step S4.2.
8. Fast Fuzzy C-Means Clustering image partition method according to claim 7, it is characterised in that Fuzzy Exponential because
Sub- m is set to 2, and iteration count initialization b2=0, it is 100 times to set highest iterations t2, and iteration ends threshold epsilon is set to 1e-5.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710558682.1A CN107368851B (en) | 2017-07-11 | 2017-07-11 | Rapid fuzzy C-means clustering image segmentation method with neighborhood selection strategy |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710558682.1A CN107368851B (en) | 2017-07-11 | 2017-07-11 | Rapid fuzzy C-means clustering image segmentation method with neighborhood selection strategy |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107368851A true CN107368851A (en) | 2017-11-21 |
CN107368851B CN107368851B (en) | 2020-12-22 |
Family
ID=60306108
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710558682.1A Active CN107368851B (en) | 2017-07-11 | 2017-07-11 | Rapid fuzzy C-means clustering image segmentation method with neighborhood selection strategy |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107368851B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109325937A (en) * | 2018-08-20 | 2019-02-12 | 中国科学院自动化研究所 | Quantity statistics method, the apparatus and system of objective body |
CN115641331A (en) * | 2022-11-18 | 2023-01-24 | 山东天意装配式建筑装备研究院有限公司 | Intelligent detection method for spraying effect of wallboard film |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060110021A1 (en) * | 2004-11-23 | 2006-05-25 | Hui Luo | Method for recognizing projection views of radiographs |
CN101251893A (en) * | 2008-03-28 | 2008-08-27 | 西安电子科技大学 | Self-adapting multi-dimension veins image segmenting method based on wavelet and average value wander |
CN101923714A (en) * | 2010-09-02 | 2010-12-22 | 西安电子科技大学 | Texture image segmentation method based on spatial weighting membership fuzzy c-mean value |
CN102750700A (en) * | 2012-06-05 | 2012-10-24 | 南京师范大学 | Fast robust fuzzy C-means image segmentation method combining neighborhood information |
CN103093478A (en) * | 2013-02-18 | 2013-05-08 | 南京航空航天大学 | Different source image rough edge test method based on rapid nuclear spatial fuzzy clustering |
CN103473739A (en) * | 2013-08-15 | 2013-12-25 | 华中科技大学 | White blood cell image accurate segmentation method and system based on support vector machine |
CN105261004A (en) * | 2015-09-10 | 2016-01-20 | 西安电子科技大学 | Mean shift and neighborhood information based fuzzy C-mean image segmentation method |
US20160217552A1 (en) * | 2015-01-22 | 2016-07-28 | Samsung Electronics Co., Ltd. | Video super-resolution by fast video segmentation for boundary accuracy control |
CN106408569A (en) * | 2016-08-29 | 2017-02-15 | 北京航空航天大学 | Brain MRI (magnetic resonance image) segmentation method based on improved fuzzy C-means clustering algorithm |
-
2017
- 2017-07-11 CN CN201710558682.1A patent/CN107368851B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060110021A1 (en) * | 2004-11-23 | 2006-05-25 | Hui Luo | Method for recognizing projection views of radiographs |
CN101251893A (en) * | 2008-03-28 | 2008-08-27 | 西安电子科技大学 | Self-adapting multi-dimension veins image segmenting method based on wavelet and average value wander |
CN101923714A (en) * | 2010-09-02 | 2010-12-22 | 西安电子科技大学 | Texture image segmentation method based on spatial weighting membership fuzzy c-mean value |
CN102750700A (en) * | 2012-06-05 | 2012-10-24 | 南京师范大学 | Fast robust fuzzy C-means image segmentation method combining neighborhood information |
CN103093478A (en) * | 2013-02-18 | 2013-05-08 | 南京航空航天大学 | Different source image rough edge test method based on rapid nuclear spatial fuzzy clustering |
CN103473739A (en) * | 2013-08-15 | 2013-12-25 | 华中科技大学 | White blood cell image accurate segmentation method and system based on support vector machine |
US20160217552A1 (en) * | 2015-01-22 | 2016-07-28 | Samsung Electronics Co., Ltd. | Video super-resolution by fast video segmentation for boundary accuracy control |
CN105261004A (en) * | 2015-09-10 | 2016-01-20 | 西安电子科技大学 | Mean shift and neighborhood information based fuzzy C-mean image segmentation method |
CN106408569A (en) * | 2016-08-29 | 2017-02-15 | 北京航空航天大学 | Brain MRI (magnetic resonance image) segmentation method based on improved fuzzy C-means clustering algorithm |
Non-Patent Citations (1)
Title |
---|
夏菁等: "结合边缘局部信息的FCM抗噪图像分割算法", 《计算机辅助设计与图形学学报》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109325937A (en) * | 2018-08-20 | 2019-02-12 | 中国科学院自动化研究所 | Quantity statistics method, the apparatus and system of objective body |
CN109325937B (en) * | 2018-08-20 | 2020-08-25 | 中国科学院自动化研究所 | Target body quantity statistical method, device and system |
CN115641331A (en) * | 2022-11-18 | 2023-01-24 | 山东天意装配式建筑装备研究院有限公司 | Intelligent detection method for spraying effect of wallboard film |
Also Published As
Publication number | Publication date |
---|---|
CN107368851B (en) | 2020-12-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105261004B (en) | Fuzzy C-mean algorithm image partition method based on average drifting and neighborhood information | |
CN106997597B (en) | It is a kind of based on have supervision conspicuousness detection method for tracking target | |
CN106778604B (en) | Pedestrian re-identification method based on matching convolutional neural network | |
CN104102919B (en) | Image classification method capable of effectively preventing convolutional neural network from being overfit | |
CN104992191B (en) | The image classification method of feature and maximum confidence path based on deep learning | |
CN105719292A (en) | Method of realizing video target tracking by adopting two-layer cascading Boosting classification algorithm | |
CN107909095A (en) | A kind of image-recognizing method based on deep learning | |
CN107563493A (en) | A kind of confrontation network algorithm of more maker convolution composographs | |
CN108399406A (en) | The method and system of Weakly supervised conspicuousness object detection based on deep learning | |
CN103186775B (en) | Based on the human motion identification method of mix description | |
CN111860267B (en) | Multichannel body-building exercise identification method based on human body skeleton joint point positions | |
CN103336992A (en) | FNN learning algorithm | |
CN107133955A (en) | A kind of collaboration conspicuousness detection method combined at many levels | |
CN105678248A (en) | Face key point alignment algorithm based on deep learning | |
CN106557579A (en) | A kind of vehicle model searching system and method based on convolutional neural networks | |
CN108647682A (en) | A kind of brand Logo detections and recognition methods based on region convolutional neural networks model | |
CN106611423A (en) | SAR image segmentation method based on ridge wave filter and deconvolution structural model | |
CN104331893A (en) | Complex image multi-threshold segmentation method | |
CN107368851A (en) | A kind of Fast Fuzzy C-Means Clustering image partition method with neighborhood choice strategy | |
CN101515369A (en) | Multiscale SAR image segmentation method based on semi-supervised learning | |
CN1992911A (en) | Target tracking method of sports video | |
CN107679539A (en) | A kind of single convolutional neural networks local message wild based on local sensing and global information integration method | |
CN106251230A (en) | A kind of community discovery method propagated based on election label | |
CN107563324A (en) | A kind of hyperspectral image classification method and device of the learning machine that transfinited based on core basis | |
CN107832753A (en) | A kind of face feature extraction method based on four value weights and multiple classification |
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 |