CN106570881A - Two-channel medical image segmentation method based on colorimetric colors and spatial nonuniformity of texture differences - Google Patents
Two-channel medical image segmentation method based on colorimetric colors and spatial nonuniformity of texture differences Download PDFInfo
- Publication number
- CN106570881A CN106570881A CN201610940333.1A CN201610940333A CN106570881A CN 106570881 A CN106570881 A CN 106570881A CN 201610940333 A CN201610940333 A CN 201610940333A CN 106570881 A CN106570881 A CN 106570881A
- Authority
- CN
- China
- Prior art keywords
- image
- texture
- information
- colourity
- space
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10068—Endoscopic image
Landscapes
- Image Analysis (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The present invention provides a two-channel medical image segmentation method based on the colorimetric colors and the spatial nonuniformity of texture differences. The method comprises the steps of S1, acquiring medical image data, subjecting the image data to preliminary screening, obtaining image data having differences, subjecting the image data having differences to space conversion to convert RGB spatial images into HIS spatial images, and obtaining a chrominance image IH and a saturation image IS from the HIS spatial images; S2, obtaining a final image IHue containing the global color information and used for image segmentation; S3, obtaining an image Itexture containing the local texture information; 4, constructing a colorimetric texture-embedded regional activity-based contour model based on the information of the image IHue and the information of the image Itexture, wherein the contour model is used for segmenting the data of a medical image having colorimetric differences and texture differences. In this way, based on the method, the automatic segmentation of interested regions is realized for specialized people. Meanwhile, the time consumption is reduced and the efficiency is improved.
Description
Technical field
The present invention relates to computer visual image segmentation field, more particularly to a kind of color and space inequality based on colourity
Even texture difference binary channels medical image（NBI）Dividing method.
Background technology
At present, ME-NBI（A kind of medical imaging modalities of Olympus research and development）Scope has become a kind of intestines for people
The effective instrument of the disease detection of gastropore.Its picture quality for amplifying high definition can be good at describing glutinous in human body intestines and stomach
The form and color of the micro-assembly robots such as the blood vessel and glandular tube of film upper strata and lower floor, is a kind of diagnosis intestines and stomach disease（Such as：Early carcinoma,
Chronic inflammation etc.）Reliable basis.But, according to the diagnosis of ME-NBI images, subjectivity is strong, extensive and beginner's threshold of classifying
Height, learning curve is steep.Meanwhile, with increasing for number of patients, amount of images is huge, but the image data of these magnanimity can only
Artificial interpretation is carried out by the scope medical worker of specialty, working strength is big, efficiency is low, consume the substantial amounts of time.Therefore, one is found
Plant reliable efficient method and split region interested automatically for professional, consume during reduction, improve efficiency.This just needs this badly
Art personnel solve corresponding technical problem.
The content of the invention
It is contemplated that at least solving technical problem present in prior art, especially innovatively propose a kind of based on color
The uneven texture difference binary channels medical image in the color of degree and space（NBI）Dividing method.
In order to realize the above-mentioned purpose of the present invention, the invention provides a kind of color and space based on colourity is uneven
The dividing method of texture difference binary channels medical image, including：
S1, gathers medical image, and preliminary screening is carried out to the view data, and acquisition possesses the picture number of otherness
According to will be provided with the view data of otherness carries out space conversion, and from rgb space image HIS space image is transformed into, empty from HIS
Between chromatic diagram is obtained in image as IHWith saturation degree image IS；
S2, using normalization rule to colourity image IHConversion is normalized, then to the chromatic diagram picture after normalization
Gassian low-pass filter is carried out, the final image I with global color information for segmentation is obtainedHue；
S3, in saturation degree image ISThe upper PIF textures by adaptive threshold describe son description textural characteristics, are carried
The image I of local grain informationtexture；
S4, by IHueAnd ItextureImage information be embedded into based on regional activity skeleton pattern（C-V models）In, build
Go out chrominance texture it is embedded based on regional activity skeleton pattern, the medical image number with colourity and texture difference for segmentation
According to.
The dividing method of the uneven texture difference binary channels medical image of the described color based on colourity and space, it is excellent
Choosing, the S1 obtains colourity image step from HIS space image to be included：
S1-1, chromatic diagram is as IhUsed as the global color information of image, colourity is to reflect space wavelength using angle
A kind of colouring information,
Wherein θ ∈ [0,2 π], Ir, Ig, IbIt is rgb space epigraph in R, G, the pixel value in channel B.
The dividing method of the uneven texture difference binary channels medical image of the described color based on colourity and space, it is excellent
Choosing, the S1 obtains saturation degree image step from HIS space image to be included：
S1-4, calculates saturation degree image Is(x), in saturation degree image IsLocal grain information, image saturation are obtained on (x)
Degree IsIt is calculated as follows：
Wherein Ir, Ig, IbBe rgb space epigraph in R, G, in channel B
Pixel value；
The dividing method of the uneven texture difference binary channels medical image of the described color based on colourity and space, it is excellent
Choosing, the S2 normalization rule step includes：
S2-1, obtains the global color frame I after calculatingHue(x),
Self-defining normalization rule is normalized to image, normalized image I' is obtained, according to equation below：
S2-2, is smoothed after being calculated to I' images using Gassian low-pass filter and reflects global color information
Chromatic diagram is as IHue(x), it is as follows：
IHue(x)=G(x)*I'h(x)
G (x) gauss low frequency filters.
The dividing method of the uneven texture difference binary channels medical image of the described color based on colourity and space, it is excellent
Choosing, the S3 includes：
S3-1, calculates local grain frame Itexture(x)；
Obtain texture image ItextureX (), is calculated as follows：
Ω ' (p)={ q ∈ N (p):| I (p)-I (q) | ＞ v }, wherein, p be image data pixel label, N (p)={ q ∈ I:
|px-qx|≤k,|py-qy|≤k, k ∈ z }, pixel point sets of the N (p) in the square region that k is yardstick centered on q, Z is
Real number set, | | the number of element in set is represented, Ω ' (p) refers to the uneven pixel point set of gray scale, and q represents center
Pixel point coordinates or label.
The dividing method of the uneven texture difference binary channels medical image of the described color based on colourity and space, it is excellent
Choosing, the S4 image informations are embedded into be included based on regional activity skeleton pattern step：
S4-1, obtains global color energy functional, global color information is embedded in and is based among regional activity skeleton pattern,
Global color energy functional computing formula is：
The inside and outside colouring information mean value of initialization curve C is represented respectively；
S4-2, obtains local grain information energy functional, and local grain information is embedded based on regional activity skeleton pattern
Among, local grain information energy Functional Calculation formula is：
WithIt is respectively image ItextureThe inside and outside mean value of initialization curve C in (x).
The dividing method of the uneven texture difference binary channels medical image of the described color based on colourity and space, it is excellent
Choosing, the medical image step of the S4 segmentations with colourity and texture difference includes：
A, initializes level set function φ0=φ0And n=0；
B, obtains the image I with global color information in colourity image channelHue(x)；
C, by self-defining PIF operators in saturation degree image channel, obtains the image with local grain information
Itexture；
D, fixed level set profile function φn, by equation below：
Digital simulation variable
F, fixed fit variationsBy equation below：
Iterate to calculate out φn+1；
F, checks whether and meets condition stop condition.If be unsatisfactory for, n=n+1 returns to D；
G, exports final medical image and φn。
The dividing method of the uneven texture difference binary channels medical image of the described color based on colourity and space, it is excellent
Choosing, also include：
Global color information and local energy information are embedded among the active contour based on region and are added regularization
, it is as follows：
λ1And λ2It is greater than 0 constant, EHFor global color information energy functional, ETIt is general for local saturation information energy
Letter, ERThe length of initial curve C is represented, is calculated as follows：
ER=μ Length (C), wherein μ >=0；
Under level set framework, will
Make following energy functional into：
Wherein, H is Sigmoid function H1,ε(z), it is as follows：
In sum, as a result of above-mentioned technical proposal, the invention has the beneficial effects as follows：
Dividing method given by the present invention has the NBI images of color and texture difference for segmentation, with adaptability
The advantages of strong and strong positioning medical science area-of-interest ability, be that professional splits region interested automatically, is consumed during reduction, is carried
High efficiency.
The additional aspect and advantage of the present invention will be set forth in part in the description, and partly will become from the following description
Obtain substantially, or recognized by the practice of the present invention.
Description of the drawings
The above-mentioned and/or additional aspect and advantage of the present invention will become from the description with reference to accompanying drawings below to embodiment
It is substantially and easy to understand, wherein：
Fig. 1 is general illustration of the present invention.
Specific embodiment
Embodiments of the invention are described below in detail, the example of the embodiment is shown in the drawings, wherein from start to finish
Same or similar label represents same or similar element or the element with same or like function.Below with reference to attached
The embodiment of figure description is exemplary, is only used for explaining the present invention, and is not considered as limiting the invention.
In describing the invention, it is to be understood that term " longitudinal direction ", " horizontal ", " on ", D score, "front", "rear",
The orientation or position relationship of the instruction such as "left", "right", " vertical ", " level ", " top ", " bottom " " interior ", " outward " is based on accompanying drawing institute
The orientation for showing or position relationship, are for only for ease of the description present invention and simplify description, rather than indicate or imply the dress of indication
Put or element must have specific orientation, with specific azimuth configuration and operation, therefore it is not intended that to the present invention limit
System.
In describing the invention, unless otherwise prescribed and limit, it should be noted that term " installation ", " connected ",
" connection " should be interpreted broadly, for example, it may be mechanically connected or electrical connection, or the connection of two element internals, can
Being to be joined directly together, it is also possible to be indirectly connected to by intermediary, for the ordinary skill in the art, can basis
Concrete condition understands the concrete meaning of above-mentioned term.
As shown in figure 1, we have proposed a kind of method of the automatic segmentation based on twin-channel Visual Feature Retrieval Process, the party
Method is to be based on two visual signatures of texture and color of image, and is embedded it among the active contour based on region, is created
Hue-texture-embedded region-based model are used to be partitioned into region interested in NBI images
（region of interest ROI）.First, we are based on chrominance information, obtain the global color energy function of image；So
Afterwards, our the PIF operators based on saturation degree image and adaptive threshold obtain the local grain energy function of image；Finally, I
Color energy and texture energy are respectively embedded in into C-V models（Based on regional activity skeleton pattern）Among（Formula 1）, build
Go out our Hue-texture-embedded region-based model, for segmentation automatically.
Wherein μ, v >=0；λ1,λ2>0 is Fixed constant.Lengt (C) is the length of initialization curve C, and inside (C) is just
The inside of beginningization curve C, outside (C) is the outside of curve C.Area (inside (C)) is the face inside initialization curve C
Product.I (x) is divided image, c1And c2It is respectively the inside and outside mean values of curve C.
1. global color functional
(1) chromatic diagram is obtained as Ih(x)
We extract the color characteristic of image from HIS color spaces, and HIS space is not only suitable for describing the color of vision
Feature also explains and reflects the internal connection between color of image and gradation of image information simultaneously.The extraction of color, Wo Menxuan
Use chromatic diagram picture（H images）As the global color information of image（Formula 2）, colourity is to reflect space wavelength using angle
A kind of colouring information.
Wherein θ ∈ [0,2 π], Ir, Ig, IbIt is rgb space epigraph in R, G, the pixel value in channel B.
（2）Obtain the global color of " clean "（Colourity）Frame IHue(x)
But, in the original chromatic diagram picture obtained from NBI images, there is the interference of too many irrelevant information.Therefore,
We obtain the image of more " clean " by 2 steps, the first step, and we are using the self-defining normalization rule of the present invention to figure
As being normalized, normalized image I' is obtainedh, according to equation below：
Second step, using Gassian low-pass filter to I'hImage is smoothed the reflection global color of acquisition " clean "
The chromatic diagram of information is as IHue(x), it is as follows：
IHue(x)=G(x)*I'h(x) （4）
G (x) gauss low frequency filters.
（3）Obtain global color energy functional
Finally, among by the embedded C-V models of global color information, global color energy functional is obtained：
The inside and outside colouring information mean value of initialization curve C is represented respectively.
2. local grain energy functional
The NBI images for splitting complexity only by global color information are cannot to obtain satisfied result, so picture
Local space relation between element can not be ignored.We obtain the local grain information of image by this local space relation.
Among the local space relation of piece image, according to certain assessment level, or the spatial neighborhood value of central pixel point is just
It is similar to its neighborhood value, on the contrary it is then dissimilar.Different assessment level, it is possible to obtain different spaces are encoded, and have different textures
Expression method.
（1）Adaptive PIF local texture description methods
We select PIF（pixel inhomogeneity factor）The inconsistent Factor Criterion of pixel is carrying out space line
Reason feature interpretation, it is as follows：
Ω ' (p)={ q ∈ N (p):| I (p)-I (q) | ＞ v }
Wherein | | the number of element in set is represented,
N(p)={q∈I:|px-qx|≤k,|py-qy|≤k, k ∈ z } in the square region that k is yardstick centered on q
Pixel point set.
The setting of threshold value v is the key for determining image texture characteristic information representation quality, but, the value of rigid regulation v
It is not well positioned to meet the extraction of NBI image texture characteristics.So, we have proposed a kind of PIF of adaptive threshold
（pixel inhomogeneity factor）, the obtaining value method of v is as follows：
1）Acquisition level, diagonal, the set of the distance between vertical and clinodiagonal direction and central pixel point, following institute
Show：
D0°I(x)={|I(i,j)-I(i+1,j)|;i=1,2,3,...,m-1,j=1,2,3,...,n}
D45°I(x)={|I(i,j)-I(i+1,j1)|;i=1,2,3,...,m-1,j=2,3,4,...,n}
D90°I(x)={|I(i,j)-I(i,j+1)|;i=1,2,3,...,m,j=1,2,3,...,n-1}
D135°I(x)={|I(i,j)-I(i-1,j-1)|;i=2,2,3,...,m,j=2,2,3,...,n} （7）
Wherein D0°I(x),D45°I(x),D90°I(x),D135°I (x) is represented respectively in image centered on pixel I (i, j)
Square Neighborhood in level, diagonal, it is vertical and the pixel and central pixel point I (i, j) in clinodiagonal direction between away from
From set；
2）Respectively calculated level, diagonal, vertically, in clinodiagonal distance set apart from frequency, obtain apart from frequency
Set C (DdI (x)) (d=0 °, 45 °, 90 °, 135 °), obtain C (DdI (x)) set in frequency be in most middle value, then obtain
Its corresponding distance value is obtained, it is as follows：
M(DdI(p))={DdI(p):min(median[C(DdI(x))])} （8）
M(DdI (p)) represent in direction is for the distance set of D, obtain frequency at point p and be in most middle distance value,
min(median[C(DdI (x))]) represent, when the distance value more than one being in corresponding to most middle frequency, go
Minimum one value in these distance values.
3）Threshold value v is calculated, it is as follows：
By combining formula（6）-（9）, obtain PIF description of adaptive threshold.
（2）Calculate saturation degree image（S images）Is(x)
In saturation degree image IsLocal grain information, image saturation I are obtained on (x)sIt is calculated as follows：
（3）Calculate local grain frame Itexture(x)
With reference to formula（6）-（10）, obtain texture image ItextureX (), is calculated as follows：
Ω ' (p)={ q ∈ N (p):|Is(p)-Is(q) | ＞ v }
（4）Obtain local grain information energy functional
WithIt is respectively image ItextureThe inside and outside mean value of initialization curve C in (x).
3.Hue-Texture-Embedded Region-Based Model（Chrominance texture it is embedded based on regional activity
Skeleton pattern）And numerical computations
（1）Model is set up
With reference to formula（5）With（12）Global color energy functional and local texture information energy functional are embedded into based on area
Among the active contour in domain and add regularization term, it is as follows：
λ1And λ2It is greater than 0 constant.
ERThe length for representing curve C is regularization term avoiding initializing reinitializing for profile C, is calculated as follows：
ER=μ Length (C) (14)
（2）Numerical computations
In order to solve the energy functional E minimization problems that we are proposed, we will solve this under level set framework
Problem.Under level set framework, we will
Make following energy functional into：
Wherein, H is Sigmoid function H1,ε(z), it is as follows：
Because this is a unconstrained optimization problem, it is possible to by respectively to being fitted 4 variablesInterleaved computation is carried out with active contour variable, it is as follows,
1）Fixed level set profile function, according toEnergy function E is minimized, such as
Under：
2) fixed fit variationsAccording to level set
Function phi minimizes energy function E, as follows：
Iterate to calculate out φn+1
T >=0 is people's work time variable, represents gradient descent direction, and generally initialization level set profile is φ0=φ
(0), δ is a smooth Dirac function, and it is H1,εDerivative, it is as follows：
In sum, the main calculation procedure of Hue-texture-embedded region-based model, following institute
Show：
Step 1：Initialization level set function φ0=φ0And n=0；
Step 2：The image I with global color information is obtained on H passagesHue(x)；
Step 3：By self-defining PIF operators in channel S, the image I with local grain information is obtainedtexture
(x)；
Step 4：By formula（17）Digital simulation variable
Step 5：By formula（18）With evolution curve φn, obtain evolution curve φn+1；
Step 6：Check whether and meet condition stop condition（Equation E reaches minimum of a value or reaches iterations）.If
It is unsatisfactory for, n=n+1 returns to step 4；
Step 7：Output result image and φn。
Effect shows
1st, 1 is tested
1）Select image
Experiment picks 12 and possesses NBI imaging models of the same race, Color and multiplication factor, but with different degrees of
Color and texture difference image：
2）Experimental result
Table 1
| Method | AFm | AFPR |
| Our model | 0.61 | 0.16 |
| C-V model | 0.52 | 0.32 |
AFm and AFPR is used to assess the quality of segmentation result, and wherein AFm values are bigger, and effect is better;AFPR values are less, effect
Fruit is better.AFm is average F-measure values, and the accurate rate of integrating representation segmentation result calls rate together with returning, and AFPR is average FPR values,
Represent the error rate of segmentation result.
Can regard as from table, our method whether error rate or accuracy rate be superior to it is traditional based on region
Dividing method.It is a kind of effective method to show our method for splitting otherness NBI image segmentation.
2. 2 are tested
1）Select image
5 different modes are selected in experiment 2, the image of coloured differently,
2）Experimental result
Table 2
Our method equally can preferably be applied to different imagings and arrange（Coloured differently, different mode, different times magnifications
Number）NBI otherness images.
（3）Experiment 3
1）Select image
Experiment 3, we pick the NBI disparity maps of 113 different peoples and different parts in image data base, at random
Picture.
2）Experimental result
Table 3
Experiment 3 is that its positioning is cured in order to prove our method for traditional method based on region segmentation
The ability of raw area-of-interest is higher.These can be as can be seen from Table 3.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means to combine specific features, structure, material or spy that the embodiment or example are described
Point is contained at least one embodiment of the present invention or example.In this manual, to the schematic representation of above-mentioned term not
Necessarily refer to identical embodiment or example.And, the specific features of description, structure, material or feature can be any
One or more embodiments or example in combine in an appropriate manner.
Although an embodiment of the present invention has been shown and described, it will be understood by those skilled in the art that：Not
These embodiments can be carried out with various changes, modification, replacement and modification in the case of the principle and objective that depart from the present invention, this
The scope of invention is limited by claim and its equivalent.
Claims (8)
1. the dividing method of the uneven texture difference binary channels medical image of a kind of color and space based on colourity, its feature
It is, including：
S1, gathers medical image, and preliminary screening is carried out to the view data, and acquisition possesses the view data of otherness, will
Possessing the view data of otherness carries out space conversion, HIS space image is transformed into from rgb space image, from HIS space image
Middle acquisition chromatic diagram is as IHWith saturation degree image IS；
S2, using normalization rule to colourity image IHConversion is normalized, then height is carried out to the chromatic diagram picture after normalization
This LPF, obtains the final image I with global color information for segmentationHue；
S3, in saturation degree image ISThe upper PIF textures by adaptive threshold describe son description textural characteristics, obtain with local
The image I of texture informationtexture；
S4, by IHueAnd ItextureImage information be embedded into based on regional activity skeleton pattern (C-V models), construct colourity
Texture it is embedded based on regional activity skeleton pattern, the medical image with colourity and texture difference for segmentation.
2. the color and space based on colourity according to claim 1 uneven texture difference binary channels medical image
Dividing method, it is characterised in that the S1 obtains colourity image step from HIS space image to be included：
S1-1, chromatic diagram is as IhUsed as the global color information of image, colourity is a kind of face for reflecting space wavelength using angle
Color information,
Wherein θ ∈ [0,2 π], Ir, Ig, IbIt is rgb space epigraph in R, G, the pixel value in channel B.
3. the color and space based on colourity according to claim 1 uneven texture difference binary channels medical image
Dividing method, it is characterised in that the S1 obtains saturation degree image step from HIS space image to be included：
S1-4, calculates saturation degree image Is(x), in saturation degree image IsLocal grain information, image saturation I are obtained on (x)s
It is calculated as follows：
Wherein Ir, Ig, IbIt is rgb space epigraph in R, G, the picture in channel B
Element value.
4. the color and space based on colourity according to claim 2 uneven texture difference binary channels medical image
Dividing method, it is characterised in that the S2 normalization rule step includes：
S2-1, obtains the global color frame I after calculatingHue(x),
Self-defining normalization rule is normalized to image, normalized image I' is obtained, according to equation below：
S2-2, is smoothed after being calculated to I' images using Gassian low-pass filter and reflects the colourity of global color information
Image IHue(x), it is as follows：
IHue(x)=G (x) * I'h(x)
G (x) gauss low frequency filters.
5. the color and space based on colourity according to claim 1 uneven texture difference binary channels medical image
Dividing method, it is characterised in that the S3 includes：
S3-1, calculates local grain frame Itexture(x)；
Obtain texture image ItextureX (), is calculated as follows：
Ω ' (p)={ q ∈ N (p):| I (p)-I (q) | ＞ v }, wherein, p be image data pixel label, N (p)={ q ∈ I:|px-
qx|≤k,|py-qy|≤k, k ∈ z }, pixel point sets of the N (p) in the square region that k is yardstick centered on q, Z is real number
Set, | | the number of element in set is represented, Ω ' (p) refers to the uneven pixel point set of gray scale, and q represents center pixel
Point coordinates or label.
6. the color and space based on colourity according to claim 1 uneven texture difference binary channels medical image
Dividing method, it is characterised in that the S4 image informations are embedded into be included based on regional activity skeleton pattern step：
S4-1, obtains global color energy functional, and global color information is embedded based among regional activity skeleton pattern, global
Color energy Functional Calculation formula is：
The inside and outside colouring information mean value of initialization curve C is represented respectively；
S4-2, obtains local grain information energy functional, local grain information is embedded in and is based among regional activity skeleton pattern,
Local grain information energy Functional Calculation formula is：
WithIt is respectively image ItextureThe inside and outside mean value of initialization curve C in (x).
7. the color and space based on colourity according to claim 1 uneven texture difference binary channels medical image
Dividing method, it is characterised in that the medical image step of the S4 segmentations with colourity and texture difference includes：
A, initializes level set function φ0=φ0And n=0；
B, obtains the image I with global color information in colourity image channelHue(x)；
C, by self-defining PIF operators in saturation degree image channel, obtains the image I with local grain informationtexture；
D, fixed level set profile function φn, by equation below：
Digital simulation variable
F, fixed fit variationsBy equation below：
Iterate to calculate out φn+1；
F, checks whether and meets condition stop condition.If be unsatisfactory for, n=n+1 returns to D；
G, exports final medical image and φn。
8. the color and space based on colourity according to claim 7 uneven texture difference binary channels medical image
Dividing method, it is characterised in that also include：
Global color information and local energy information are embedded among the active contour based on region and are added regularization term, such as
Shown in lower：
λ1And λ2It is greater than 0 constant, EHFor global color information energy functional, ETFor local saturation information energy functional, ER
The length of initial curve C is represented, is calculated as follows：
ER=μ Length (C), wherein μ >=0；
Under level set framework, will
Make following energy functional into：
Wherein, H is Sigmoid function H1,ε(z), it is as follows：
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610940333.1A CN106570881B (en) | 2016-10-25 | 2016-10-25 | The dividing method of the non-uniform texture difference binary channels medical image of color and space based on coloration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610940333.1A CN106570881B (en) | 2016-10-25 | 2016-10-25 | The dividing method of the non-uniform texture difference binary channels medical image of color and space based on coloration |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106570881A true CN106570881A (en) | 2017-04-19 |
| CN106570881B CN106570881B (en) | 2019-08-13 |
Family
ID=58534714
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610940333.1A Active CN106570881B (en) | 2016-10-25 | 2016-10-25 | The dividing method of the non-uniform texture difference binary channels medical image of color and space based on coloration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106570881B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107220673A (en) * | 2017-06-06 | 2017-09-29 | 滁州市天达汽车部件有限公司 | A kind of bamboo cane method for sorting colors based on KNN algorithms |
| CN112257711A (en) * | 2020-10-26 | 2021-01-22 | 哈尔滨市科佳通用机电股份有限公司 | Method for detecting damage fault of railway wagon floor |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1852392A (en) * | 2006-05-11 | 2006-10-25 | 上海交通大学 | Printing net-point-image dividing method based on moveable contour |
| US20100172576A1 (en) * | 2008-07-11 | 2010-07-08 | Ilia Goldfarb | Color Analyzer And Calibration Tool |
| CN102354396A (en) * | 2011-09-23 | 2012-02-15 | 清华大学深圳研究生院 | Method for segmenting image with non-uniform gray scale based on level set function |
| CN103093473A (en) * | 2013-01-25 | 2013-05-08 | 北京理工大学 | Multi-target picture segmentation based on level set |
| JP2014023104A (en) * | 2012-07-23 | 2014-02-03 | Nikon Corp | Image processing device, imaging devices, and program |
| CN105844625A (en) * | 2016-03-18 | 2016-08-10 | 常州大学 | Movable profile image segmentation method fusing edge and area |
-
2016
- 2016-10-25 CN CN201610940333.1A patent/CN106570881B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1852392A (en) * | 2006-05-11 | 2006-10-25 | 上海交通大学 | Printing net-point-image dividing method based on moveable contour |
| US20100172576A1 (en) * | 2008-07-11 | 2010-07-08 | Ilia Goldfarb | Color Analyzer And Calibration Tool |
| CN102354396A (en) * | 2011-09-23 | 2012-02-15 | 清华大学深圳研究生院 | Method for segmenting image with non-uniform gray scale based on level set function |
| JP2014023104A (en) * | 2012-07-23 | 2014-02-03 | Nikon Corp | Image processing device, imaging devices, and program |
| CN103093473A (en) * | 2013-01-25 | 2013-05-08 | 北京理工大学 | Multi-target picture segmentation based on level set |
| CN105844625A (en) * | 2016-03-18 | 2016-08-10 | 常州大学 | Movable profile image segmentation method fusing edge and area |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107220673A (en) * | 2017-06-06 | 2017-09-29 | 滁州市天达汽车部件有限公司 | A kind of bamboo cane method for sorting colors based on KNN algorithms |
| CN107220673B (en) * | 2017-06-06 | 2020-05-01 | 安徽天达汽车制造有限公司 | KNN algorithm-based bamboo strip color classification method |
| CN112257711A (en) * | 2020-10-26 | 2021-01-22 | 哈尔滨市科佳通用机电股份有限公司 | Method for detecting damage fault of railway wagon floor |
| CN112257711B (en) * | 2020-10-26 | 2021-04-09 | 哈尔滨市科佳通用机电股份有限公司 | Method for detecting damage fault of railway wagon floor |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106570881B (en) | 2019-08-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107230206B (en) | Multi-mode data-based 3D pulmonary nodule segmentation method for hyper-voxel sequence lung image | |
| US10878574B2 (en) | 3D quantitative analysis of retinal layers with deep learning | |
| Liao et al. | Automatic liver segmentation from abdominal CT volumes using graph cuts and border marching | |
| CN108961261B (en) | Optic disk region OCT image hierarchy segmentation method based on space continuity constraint | |
| SG194407A1 (en) | A method and system for determining a stage of fibrosis in a liver | |
| CN105657580B (en) | A kind of capsule endoscope video abstraction generating method | |
| CN106600596A (en) | Dynamic display method and system for heart cavity | |
| JP5771442B2 (en) | Image processing apparatus, image processing method, and image processing program | |
| CN108932536A (en) | Human face posture method for reconstructing based on deep neural network | |
| CN113012173A (en) | Heart segmentation model and pathology classification model training, heart segmentation and pathology classification method and device based on cardiac MRI | |
| Celebi et al. | Fast and accurate border detection in dermoscopy images using statistical region merging | |
| Cheng et al. | Watershed-presegmented snake for boundary detection and tracking of left ventricle in echocardiographic images | |
| CN110751680A (en) | Image processing method with fast alignment algorithm | |
| CN110969635B (en) | Robust rapid image segmentation method based on priori constraint level set frame | |
| Senthilkumaran et al. | Brain image segmentation | |
| Mendonca et al. | Comparison of segmentation methods for automatic diagnosis of dermoscopy images | |
| CN115393239A (en) | Multi-mode fundus image registration and fusion method and system | |
| CN106570881A (en) | Two-channel medical image segmentation method based on colorimetric colors and spatial nonuniformity of texture differences | |
| Sookpotharom | Border detection of skin lesion images based on fuzzy C-means thresholding | |
| CN104751457A (en) | Novel variational energy based liver partition method | |
| CN110517766A (en) | Identify the method and device of encephalatrophy | |
| Kim et al. | Tongue diagnosis method for extraction of effective region and classification of tongue coating | |
| Skalski et al. | Left ventricle USG image segmentation using active contour model | |
| CN108876789B (en) | Continuous graph cutting method based on combination of super-pixel and neighborhood block features | |
| IT201900007806A1 (en) | Computerized method of classifying a mass of an organ as a cyst |
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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20211109 Address after: 401120 1-1, 2-1, 3-1, building 5, No. 18, Cuiping Lane 2, Huixing street, Yubei District, Chongqing Patentee after: Chongqing Jinshan Medical Technology Research Institute Co.,Ltd. Address before: 401120 1 office buildings, Jinshan International Industrial City, 18 of Nei sang Road, Hui Xing street, Yubei District, Chongqing. Patentee before: CHONGQING JINSHAN MEDICAL APPLIANCE Co.,Ltd. |