CN106340023A - Image segmentation method and image segmentation device - Google Patents
Image segmentation method and image segmentation device Download PDFInfo
- Publication number
- CN106340023A CN106340023A CN201610702051.8A CN201610702051A CN106340023A CN 106340023 A CN106340023 A CN 106340023A CN 201610702051 A CN201610702051 A CN 201610702051A CN 106340023 A CN106340023 A CN 106340023A
- Authority
- CN
- China
- Prior art keywords
- pixel
- image
- area
- target area
- brightness value
- 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
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
- G06T2207/30201—Face
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention puts forward an image segmentation method comprising the steps of acquiring a to-be-segmented original target image, extracting the geometrical characteristics of a to-be-segmented region in the original target image, acquiring the color characteristics of each pixel in the original target image, clustering the pixels in the original target image according to the extracted geometrical characteristics and the color characteristics of each pixel, and segmenting the original target image according to the clustering result. The pixels in the original target image are clustered according to both the geometrical characteristics and the color characteristics of each pixel. Thus, even if the color characteristics of the to-be-segmented region are not obvious, the region can still be segmented precisely due to the limitation of geometrical characteristics. In addition, the invention puts forward an image segmentation device.
Description
Technical field
The present invention relates to image processing field, more particularly to a kind of method and apparatus of image segmentation.
Background technology
With the development of image technique, how segmentation is carried out to image and seem more and more important.Traditional area to image
Domain carries out splitting mainly two methods, and a kind of is dividing method based on color characteristic, and one kind is based on edge feature and energy
The dividing method of amount functional.Wherein, the dividing method based on color characteristic has higher requirement to the depth of color, if segmentation
The color characteristic of target is inconspicuous, can cause to split unsuccessfully.Such as, when the very slight color of lip is almost consistent with the face colour of skin
When, just successful division cannot go out lip region with this dividing method.Energy functional is exactly substantially first to set up description set of regions
Close the parameter expression of feature, change the shape in region by regulation parameter, when the energy functional of definition takes minimum of a value, table
Reach the edges of regions that the geometric figure represented by formula gears to actual circumstances completely, its shortcoming is that the geometric properties in region are more complicated
In the case of resume parameter expression relatively difficult, and parameter is numerous, and parameter is numerous, and itself energy function to be led to solve minimum again
The difficulty of value is very big, and the operational efficiency of algorithm also becomes very low, and its curve described is usually continuously and smooth
, and the border of actually some cutting objects itself is not smooth, so region segmentation effect is bad.
Content of the invention
Based on this it is necessary to be directed to the problems referred to above, provide a kind of method and apparatus of more accurate image segmentation
A kind of method of image segmentation, methods described includes:
Obtain original target image to be split；
Extract the geometric properties in region to be split in described original target image；
Obtain the color characteristic of each pixel in described original target image；
Color characteristic according to the described geometric properties extracting and each pixel described is to each in original target image
Individual pixel is clustered；
According to the result of cluster, described original target image is split.
A kind of device of image segmentation, described device includes:
Image collection module, for obtaining original target image to be split；
Extraction module, for extracting the geometric properties in region to be split in described original target image；
Color characteristic acquisition module, for obtaining the color characteristic of each pixel in described original target image；
First cluster module, for according to extract described geometric properties and each pixel described color characteristic to former
Each pixel in beginning target image is clustered；
Segmentation module, for being split described original target image according to the result of cluster.
The method and apparatus of above-mentioned image segmentation, by obtaining original target image to be split, extracts original object figure
The geometric properties in region to be split in picture, obtain the color characteristic of each pixel in original target image, several according to extract
The color characteristic of what feature and each pixel clusters to each pixel in original target image.Above-mentioned image segmentation
Method, by combining the color characteristic of geometric properties and each pixel, each pixel in original target image is clicked through
Row cluster, accordingly even when the color characteristic in region to be split is inconspicuous, the restriction due to geometric properties also can be accurately by region
Split.Therefore, by combining more accurately region to be entered by the geometric properties in region to be split and color characteristic
Row segmentation.
Brief description
Fig. 1 is the internal structure schematic diagram of terminal in an embodiment；
Fig. 2 is the method flow diagram of image segmentation in an embodiment；
Fig. 3 is the schematic diagram of the point extracting Face geometric eigenvector in an embodiment；
Fig. 4 is the method stream according to the geometric properties extracting and color characteristic, pixel being clustered in an embodiment
Cheng Tu；
Fig. 5 a is the schematic diagram of the original target image in an embodiment；
Fig. 5 b is that in an embodiment, lip-region is filled and carried out the schematic diagram after Fuzzy Processing；
Fig. 6 is the method flow diagram according to color characteristic and brightness value, pixel being clustered in an embodiment；
Fig. 7 is the method flow diagram that in an embodiment, image is carried out with Fuzzy Processing；
Fig. 8 is the structured flowchart of the device of image segmentation in an embodiment；
Fig. 9 is the structured flowchart of the first cluster module in an embodiment；
Figure 10 is the structured flowchart of the second cluster module in an embodiment；
Figure 11 is the structured flowchart of fuzzy processing module in an embodiment.
Specific embodiment
In order that the objects, technical solutions and advantages of the present invention become more apparent, below in conjunction with drawings and Examples, right
The present invention is further elaborated.It should be appreciated that specific embodiment described herein is only in order to explain the present invention, and
It is not used in the restriction present invention.
As shown in figure 1, in one embodiment, the internal structure of terminal 100 is as shown in figure 1, include by system bus
Processor, built-in storage, non-volatile memory medium, network interface, image collecting device, display screen and the input dress connecting
Put.Wherein, the non-volatile memory medium of terminal 100 is stored with operating system, also includes a kind of device of image segmentation, this figure
As the device of segmentation is used for realizing a kind of method of image segmentation.This processor is used for providing calculating and control ability, supports whole
The operation of individual terminal.The operation of the device for the image segmentation in non-volatile memory medium for the built-in storage in terminal provides ring
, there is computer-readable instruction in border in this built-in storage, and this computer-readable instruction is when executed by, and can make processor
Execute a kind of method of image segmentation.Network interface is used for being connected to network and is communicated.Image collecting device is used for image
Collection, such as carries out the typing of image.The display screen of terminal can be LCDs or electric ink display screen etc., input
Device can be the button of setting, trace ball or touch-control on the touch layer or electronic equipment casing covering on display screen
Plate or external keyboard, Trackpad or mouse etc..This terminal can be mobile phone, panel computer, notebook computer, platform
Formula computer etc..It will be understood by those skilled in the art that the structure shown in Fig. 1, the only portion related to application scheme
The block diagram of separation structure, does not constitute the restriction of the terminal that application scheme is applied thereon, and specific terminal can include
Ratio part more or less of shown in figure, or combine some parts, or there are different part arrangements.
As shown in Fig. 2 in one embodiment it is proposed that a kind of method of image segmentation, the method includes:
Step 202, obtains original target image to be split.
Specifically, original target image can be coloured image or gray level image.Original target image is carried out
The premise of segmentation is that the region to be split in image has certain geometry restriction.
Step 204, extracts the geometric properties in region to be split in original target image.
In the present embodiment, the geometry spy in region to be split is extracted according to the geometry that region to be split has in itself
Levy.The geometric properties extracting region to be split have a variety of methods, can take different methods for different images.Its
In, for facial image, the point of description Face geometric eigenvector, the feature that then will extract can be extracted using Face detection technology
Point couples together and is the formation of corresponding geometric properties.Taking extract the lip in face picture as a example, Fig. 3 is in an embodiment
Extract the schematic diagram of the point of Face geometric eigenvector, wherein, include the point (65-82) of description lip-region geometric properties, will retouch
The characteristic point stating lip couples together the geometric properties just constituting this lip-region.For inhuman face image, can adopt and appoint
What computer picture algorithm, to extract the geometric properties in region to be split, does not limit to the algorithm extracting geometric properties here
System.
Step 206, obtains the color characteristic of each pixel in original target image.
Specifically, image is made up of pixel one by one, and each pixel corresponds to a color characteristic.Original mesh
Logo image can be coloured image can also be gray level image, if image is coloured image, it is possible to use multiple color space
Component is combined, and is one such as using the cbcr component combination that three components of lab color space add ycbcr color space
The individual color characteristic vector (l, a, b, cb, cr) comprising five components, wherein, the l in lab represents brightness, and a represents from carmetta
Arrive the scope of green, b represents from yellow to blue scope；Cb in ycbcr represents the concentration excursion amount of blueness, and cr represents red
The concentration excursion amount of color.If image is gray level image, then color characteristic is only just permissible with half-tone information.Obtain original object
In image, the color characteristic of each pixel is it is simply that the color characteristic vector representing pixel will be obtained.
Step 208, the color characteristic according to the geometric properties extracting and each pixel is to each in original target image
Individual pixel is clustered.
In the present embodiment, first, determine target area to be split in original target image according to the geometric properties extracting
Domain and nontarget area, because the target area obtaining and nontarget area according to geometric properties division might not be accurate, have
When also can there is relatively large deviation, so needing color characteristic further combined with each pixel to the target area and non-dividing
Target area is corrected adjustment and obtains more accurately dividing.Specifically, determined in original target image according to geometric properties and treat
Behind the target area of segmentation and nontarget area, the different color in target area and nontarget area is filled with to distinguish
Target area and nontarget area, the image after filling are carried out Fuzzy Processing, extract in the image after carrying out Fuzzy Processing
For characterizing the parameter of different colours feature, such as, the bright of each pixel in the image after carrying out Fuzzy Processing can be extracted
Angle value, then the color characteristic according to each pixel and corresponding brightness value click through to each pixel in original target image
Row cluster, it should be noted that being not limited to the extraction to brightness value here, can also be that other can be with reaction color feature
Parameter, such as, chroma, saturation degree etc..
Original target image is split by step 210 according to the result of cluster.
In the present embodiment, the result obtaining through cluster is the probability that each pixel is under the jurisdiction of target area, obtains
After each pixel is under the jurisdiction of the probability of target area, binary conversion treatment can be carried out to it, that is, will be greater than predetermined probabilities value
Pixel is divided into target area, and the pixel less than or equal to predetermined probabilities value is divided into nontarget area.
In the present embodiment, by obtaining original target image to be split, extract area to be split in original target image
The geometric properties in domain, obtain original target image in each pixel color characteristic, according to extract geometric properties and each
The color characteristic of pixel clusters to each pixel in original target image.The method of above-mentioned image segmentation, passes through
The color characteristic of geometric properties and each pixel is combined each pixel in original target image is clustered, so
Even if the color characteristic in region to be split is inconspicuous, region also can accurately be split by the restriction due to geometric properties.Cause
This, by combining the geometric properties in region to be split and color characteristic and can more accurately region be split.
As shown in figure 4, in one embodiment, the color characteristic according to the geometric properties extracting and each pixel is to former
The step 208 that each pixel in beginning target image is clustered includes:
Step 402, the geometric properties according to extracting determine target area to be split in original target image and non-targeted
Region.
In the present embodiment, after being extracted the geometric properties in region to be split, according to the geometry in the region to be split extracted
Original target image has been divided into two parts by feature, and one is target area, and one is nontarget area.Wherein, target
Region can be one or multiple.Such as, taking face picture as a example, as shown in figure 3, can be according to the face extracting
The geometric properties of (eyebrow, eye, nose, lip, face's outline) are by the face in face simultaneously as target area, other parts conduct
Nontarget area.Can certainly only using certain region as target area, such as, only extract the geometric properties of lip, by lip
As target area, other are as nontarget area in portion.
Step 404, the target area color different with nontarget area brightness value is filled with to distinguish target area
Domain and nontarget area.
In the present embodiment, after having determined target area to be split in original target image and nontarget area, by mesh
The mark region color different with nontarget area brightness value is filled with to distinguish this target area and nontarget area.In order to
Preferably target area and nontarget area are distinguished, the brightness value difference of the Fill Color of use is the bigger the better.Preferably,
Target area can be filled with white (brightness value is for 255), nontarget area is filled with black (brightness value is for 0)
Or target area is filled with black, nontarget area is filled with white.
Step 406, the image after filling is carried out Fuzzy Processing, obtains each pixel in the image after carrying out Fuzzy Processing
The corresponding brightness value of point.
In the present embodiment, because the target area that the geometric properties in the region to be split according to extraction obtain might not
Accurately, there is relatively large deviation sometimes, so needing to carry out Fuzzy Processing to the image after filling, subsequently in conjunction with color characteristic
The target area and nontarget area dividing is corrected by adjustment and obtains more accurately dividing.Specifically, mould is carried out to image
Paste is processed it is necessary first to determine blur radius, and the size of blur radius depends on extracting the deviation of region geometry feature to be split
Size, deviation then should increase greatly blur radius, and deviation is little then to reduce blur radius.The size of blur radius can be rule of thumb
Value is pre-set, and after having determined the size of blur radius, using Gaussian Blur algorithm, the image after filling can be carried out
Fuzzy Processing.Fig. 5 a is the original target image in an embodiment, Fig. 5 b be lip-region is filled with after obscured
Schematic diagram after process, wherein, the lip-region of extraction is filled with white, and other nontarget areas are carried out with black
Filling.Image after filling is carried out after Fuzzy Processing, calculates the corresponding brightness value of each pixel, the calculating of brightness value can
Using existing computational methods, any restriction is not done to the calculating of brightness value here, for the ease of the calculating of brightness value, preferably
Can select with black and white, target area and nontarget area are filled with, it is right that the calculating of such brightness value can be reduced to
The calculating of gray value.
Step 408, the color characteristic according to each pixel and corresponding brightness value are to each in original target image
Pixel is clustered.
In the present embodiment, every pixel has unique coordinate in the picture, by coordinate same before and after image procossing
The color characteristic of the corresponding pixel in position and brightness value merge, and obtain the characteristic vector of a multidimensional representing this pixel.
Such as, if in original target image, each pixel is corresponding is 5 latitudes color characteristic vector (l, a, b, cb, cr), then
With carry out each pixel corresponding brightness value b (brightness, brightness) after Fuzzy Processing combination after reformed into one
Individual 6 latitude characteristic vectors (l, a, b, cb, cr, b).According to each pixel corresponding multidimensional characteristic vector obtaining to original object
Pixel in image is clustered.Specifically, first, according to each pixel corresponding brightness value b in the image after processing
Determine the corresponding initial degree of membership of each pixel；Secondly, according to the corresponding initial degree of membership of each pixel and each pixel
The corresponding multidimensional characteristic Vector operation target area of point and the initial centered value of nontarget area；Finally, according to initial degree of membership
With initial centered value and each pixel corresponding multidimensional characteristic vector, the pixel in original target image is clustered
Obtain the probability that each pixel is under the jurisdiction of target area.The probability that each pixel according to obtaining is under the jurisdiction of target area enters
The segmentation of row image.
In the present embodiment, carry out the division in region by extracting the geometric properties in region to be split before this, then root again
To correct the target area marking off by geometric properties according to color characteristic, accordingly even when the color characteristic of the target area of segmentation
Inconspicuous, target area also can accurately be divided out by the restriction due to geometric properties, by by the geometry in region to be split
Feature and color characteristic combine and can more accurately region be split.
As shown in fig. 6, in one embodiment, color characteristic according to each pixel and corresponding brightness value are to original
The step 408 that each pixel in target image is clustered includes:
Step 408a, according to the corresponding brightness value of each pixel determine respectively each pixel be under the jurisdiction of target area and
The initial degree of membership of nontarget area.
In the present embodiment, degree of membership belongs to the concept in fuzzy evaluation functions, specifically, if to arbitrary in domain u
Element, has number a (x) ∈ [0,1] to correspond to therewith, then a is called the fuzzy set on u, and a (x) becomes the degree of membership to a for the x.When
When x changes in u, a (x) is exactly a function, the referred to as membership function of a.Degree of membership a (x) is closer to 1, represents that x belongs to a
Degree higher, degree of membership a (x) is closer to 0, represent x belong to a degree lower, that is, that uses interval [0,1] is subordinate to letter
Number a (x) represents the degree height belonging to a.It is necessary first to determine that each pixel is under the jurisdiction of target area before being clustered
Initial degree of membership and each pixel be under the jurisdiction of the initial degree of membership of nontarget area.The determination of initial degree of membership is to make
Finally belong to the probability of target area for each pixel of calculation of initial value.It is under the jurisdiction of the probability (degree of membership) of target area and be subordinate to
Probability (degree of membership) sum belonging to nontarget area is 1.Assume that the degree of membership that pixel is under the jurisdiction of target area is a (xi),
The degree of membership being under the jurisdiction of nontarget area is b (xi), wherein, a (xi)+b(xi)=1, xiRepresent the data that ith measurement arrives.Just
Beginning degree of membership determines according to brightness value.Specifically, in one embodiment, the pixel that will be greater than predetermined luminance value is under the jurisdiction of mesh
The initial degree of membership in mark region is set to 1, the pixel of no more than predetermined luminance value is under the jurisdiction of the initial degree of membership of target area
It is set to 0, then the initial degree of membership being under the jurisdiction of nontarget area more than the pixel of predetermined luminance value is 0, no more than default bright
The initial degree of membership that the pixel of angle value is under the jurisdiction of target area is 1.Specifically, with reference to Fig. 5 b, xiBelong to the first of target area
The computational methods of beginning degree of membership are: if xiThe brightness value on position in figure 5b is more than 128, then initial degree of membership is 1, little
It is 0 in degree of membership initial equal to 128.xiThe initial degree of membership belonging to nontarget area just and belongs to the initial of target area
Degree of membership is consistent, and that is, brightness value is more than 128, and initial degree of membership is 0, is 1 less than or equal to 128 initial degrees of membership.
Step 408b, the color characteristic according to each pixel and corresponding brightness value determine target area and non-mesh respectively
The initial centered value in mark region.
In the present embodiment, before being clustered, the cluster centre first determining target area and nontarget area is needed to be
Initial centered value.Calculate initial centered value, average weighted computational methods can be adopted.Specifically, formula can be passed throughIt is calculated, wherein, cjRepresent the cluster centre of class j；xiRepresent what ith measurement arrived
Multidimensional data；uijIt is xiBelong to the degree of membership of classification j；M is a flexible parameter of control algolithm, typically takes m=2；N represents total
Data amount check.Wherein, characteristic xiIt is exactly each pixel corresponding multidimensional characteristic vector in the present embodiment (by color
Feature and brightness value composition)；uijEach pixel exactly obtaining belongs to the initial degree of membership of classification j.In one embodiment,
With reference to Fig. 5 b, the cluster centre of target area is that the computational methods of initial centered value are: all brightness values are equal to 255 in figure 5b
Pixel to calculate the draw value of its multidimensional characteristic vector in original target image (Fig. 5 a) be initial centered value c1.Equally
, in nontarget area, the pixel equal to 0 for all brightness values calculates its multidimensional characteristic in original target image in figure 5b
Draw value c of vector2.c1And c2Be respectively the cluster centre of target area and nontarget area be initial centered value.
Step 408c, the color characteristic according to initial degree of membership and initial centered value and each pixel and corresponding bright
Angle value carries out clustering the probability obtaining that each pixel is under the jurisdiction of target area to the pixel in original target image.
In the present embodiment it is determined that each pixel be under the jurisdiction of target area and nontarget area initial degree of membership and
After the cluster centre of target area and nontarget area, using initial degree of membership and initial centered value as initial parameter, by each
The color characteristic of pixel and corresponding brightness value are combined as a multidimensional characteristic vector as input variable, are then calculated using cluster
Method carries out cluster calculation.In one embodiment, can be iterated being calculated each pixel person in servitude using fcm clustering algorithm
Belong to the final degree of membership of target area, carry out the segmentation in region according to this final degree of membership (probability).Specifically, by formulaWith
It is updated iteration and make object functionReach minimum.Wherein, m is more than 1
Real number, typically takes m=2, uijIt is xiBelong to the degree of membership of classification j, cjRepresent the cluster centre of class j, xiRepresent what ith measurement arrived
Multidimensional data, | | * | | represents the pixel degree of arbitrary measurement data and cluster centre, and k represents the span of c, and c represents classification number
Amount.First, the initial degree of membership by determiningSubstitute into above-mentioned formula (1) and calculate cluster centre cj, then substitute into above-mentioned again
Formula (2) calculates the degree of membership of a new roundBy so continuous iteration until
When, iteration stopping.Wherein 0 < ε < 1 is iteration ends parameter, and n is iteration wheel number, j in this processmConverge to one minimum
Value.The final degree of membership that each pixel is under the jurisdiction of target area can be obtained through iterative calculation, according to this final degree of membership
(probability) carries out the segmentation of image.
As shown in fig. 7, in one embodiment, the image after filling is carried out Fuzzy Processing, calculates and carry out Fuzzy Processing
In image afterwards, the step 406 of the brightness value of each pixel includes:
Step 406a, obtains default blur radius, carries out Fuzzy Processing according to blur radius to the image after filling.
In the present embodiment, after the target area in original target image and nontarget area being filled with, it is right to need
Image after filling carries out Fuzzy Processing, this is because the target area being divided by geometric properties and nontarget area are not necessarily
Accurately in fact it could happen that deviation, so needing to carry out Fuzzy Processing to the image after filling, subsequently in conjunction with color characteristic to division
Target area and nontarget area be corrected adjustment and obtain more accurately dividing.It is necessary first to really before carrying out Fuzzy Processing
Determine blur radius, the size of blur radius can be pre-set, and the size of the blur radius pre-setting depends on
Deviation size.Deviation size is an empirical value, can evaluate the deviation of the algorithm extracting geometric properties through multiple test
Scope, then can determine blur radius according to deviation range, and blur radius typically take the value more than maximum deviation.Such as deviation
Scope is 0-10, then blur radius may be greater than 10 value, and being so easy to subsequently can be more accurately to target area
Divided.
Step 406b, calculates the corresponding brightness value of each pixel in the image after carrying out Fuzzy Processing.
In the present embodiment, in order to make a distinction target area and nontarget area, respectively by target area and non-mesh
Brightness value different color in mark region is filled with, and then carries out Fuzzy Processing to the image after filling, thus obtains
One new target image, calculates the corresponding brightness value of each pixel in this new target image.For the ease of brightness value
Calculate, typically target area is carried out with white filling, nontarget area carries out filled black, so not only can preferably distinguish
Target area and nontarget area, can also simplify the calculating of brightness value, because only needing when being filled and calculated brightness value by black and white
Gray count to be carried out can be obtained by the corresponding brightness value of each pixel.If be filled with colour, then calculate bright
Need during angle value the value of three kinds of colors is weighted with average computation just to obtain the corresponding brightness value of pixel.
In one embodiment, region is carried out according to the probability that each pixel that cluster obtains is under the jurisdiction of target area
Segmentation, wherein, the pixel that will be greater than predetermined probabilities value is divided into target area, by the pixel less than or equal to predetermined probabilities value
Point is divided into nontarget area.
In the present embodiment, obtain degree of membership that each pixel belongs to target area at the end of cluster iteration i.e.
Each pixel belongs to the probability of target area eventually.The probability obtaining is carried out a binary conversion treatment, that is, according to probabilistic determination
Whether each pixel belongs to target area.Specifically, a probable value can be set, and such as probable value is set to 0.6, by probability
Pixel more than 0.6 is divided into target area, and the pixel that probable value is less than or equal to 0.6 is divided into nontarget area.
As shown in figure 8, in one embodiment it is proposed that a kind of device of image segmentation, this device includes:
Image collection module 802, for obtaining original target image to be split.
Extraction module 804, for extracting the geometric properties in region to be split in original target image.
Color characteristic acquisition module 806, for obtaining the color characteristic of each pixel in original target image.
First cluster module 808, for according to extract geometric properties and each pixel color characteristic to original mesh
Each pixel in logo image is clustered.
Segmentation module 810, for being split original target image according to the result of cluster.
As shown in figure 9, in one embodiment, the first cluster module 808 includes:
Determining module 902, for determining target area to be split in original target image according to the geometric properties extracting
And nontarget area.
Filling module 904, is filled with for the color that target area is different with nontarget area brightness value with area
Partial objectives for region and nontarget area.
Fuzzy Processing module 906, for the image after filling is carried out Fuzzy Processing, calculates the figure after carrying out Fuzzy Processing
The corresponding brightness value of each pixel in picture.
Second cluster module 908, for the color characteristic according to each pixel and corresponding brightness value to original object
Each pixel in image is clustered.
As shown in Figure 10, in one embodiment, the second cluster module 908 includes:
Initial degree of membership determining module 908a, for determining each pixel respectively according to the corresponding brightness value of each pixel
Point is under the jurisdiction of the initial degree of membership of target area and nontarget area.
Initial centered value determining module 908b, for the color characteristic according to each pixel and corresponding brightness value difference
Determine the initial centered value of target area and nontarget area.
3rd cluster module 908c, special for the color according to initial degree of membership and initial centered value and each pixel
Corresponding brightness value of seeking peace carries out cluster and obtains each pixel being under the jurisdiction of target area to the pixel in original target image
Probability.
As shown in figure 11, Fuzzy Processing module 906 includes:
Blur radius acquisition module 906a, for obtaining default blur radius, according to blur radius to the figure after filling
As carrying out Fuzzy Processing.
Brightness value computing module 906b, carries out the corresponding brightness of each pixel in the image after Fuzzy Processing for calculating
Value.
In one embodiment, each pixel that segmentation module is additionally operable to according to cluster obtains is under the jurisdiction of target area
Probability carries out the segmentation in region, and wherein, the pixel that will be greater than predetermined probabilities value is divided into target area, will be less than or equal to pre-
If the pixel of probable value is divided into nontarget area.
One of ordinary skill in the art will appreciate that realizing all or part of flow process in above-described embodiment method, it is permissible
Instruct related hardware to complete by computer program, this computer program can be stored in an embodied on computer readable storage and be situated between
In matter, this program is upon execution, it may include as the flow process of the embodiment of above-mentioned each method.Wherein, aforesaid storage medium can be
The non-volatile memory mediums such as magnetic disc, CD, read-only memory (read-only memory, rom), or random storage note
Recall body (random access memory, ram) etc..
Embodiment described above only have expressed the several embodiments of the present invention, and its description is more concrete and detailed, but simultaneously
Therefore the restriction to the scope of the claims of the present invention can not be interpreted as.It should be pointed out that for those of ordinary skill in the art
For, without departing from the inventive concept of the premise, some deformation can also be made and improve, these broadly fall into the guarantor of the present invention
Shield scope.Therefore, the protection domain of patent of the present invention should be defined by claims.
Claims (10)
1. a kind of method of image segmentation, methods described includes:
Obtain original target image to be split；
Extract the geometric properties in region to be split in described original target image；
Obtain the color characteristic of each pixel in described original target image；
Color characteristic according to the described geometric properties extracting and each pixel described is to each picture in original target image
Vegetarian refreshments is clustered；
According to the result of cluster, described original target image is split.
2. method according to claim 1 it is characterised in that described according to extract described geometric properties and described each
The step that the color characteristic of pixel is clustered to each pixel in original target image includes:
Described geometric properties according to extracting determine target area to be split in described original target image and nontarget area；
The described target area color different with described nontarget area brightness value is filled with to distinguish described target area
Domain and described nontarget area；
Image after filling is carried out Fuzzy Processing, calculates the corresponding brightness of each pixel in the image after carrying out Fuzzy Processing
Value；
Color characteristic according to each pixel described and corresponding described brightness value are to each in described original target image
Pixel is clustered.
3. method according to claim 2 is it is characterised in that the color characteristic of each pixel described in described basis and right
The step that the described brightness value answered is clustered to each pixel in described original target image includes:
Determine that each pixel is under the jurisdiction of described target area and described respectively according to the corresponding brightness value of each pixel described
The initial degree of membership of nontarget area；
Color characteristic according to each pixel described and corresponding described brightness value determine target area and non-target area respectively
The initial centered value in domain；
Color characteristic according to described initial degree of membership and described initial centered value and each pixel described and corresponding institute
State brightness value the pixel in original target image is carried out clustering and obtain each pixel and be under the jurisdiction of the general of described target area
Rate.
4. method according to claim 2, it is characterised in that described carry out Fuzzy Processing by the image after filling, obtains
The step carrying out the brightness value of each pixel in the image after Fuzzy Processing includes:
Obtain default blur radius, according to described blur radius, Fuzzy Processing is carried out to the image after filling；
Calculate the corresponding brightness value of each pixel in the image after carrying out described Fuzzy Processing.
5. method according to claim 1 is it is characterised in that the described result according to cluster is by described original target image
The step split includes:
Carry out the segmentation in region according to the probability that each pixel that cluster obtains is under the jurisdiction of described target area, wherein, will be big
Pixel in predetermined probabilities value is divided into target area, and the pixel less than or equal to predetermined probabilities value is divided into non-targeted
Region.
6. a kind of device of image segmentation is it is characterised in that described device includes:
Image collection module, for obtaining original target image to be split；
Extraction module, for extracting the geometric properties in region to be split in described original target image；
Color characteristic acquisition module, for obtaining the color characteristic of each pixel in described original target image；
First cluster module, for according to extract described geometric properties and each pixel described color characteristic to original mesh
Each pixel in logo image is clustered；
Segmentation module, for being split described original target image according to the result of cluster.
7. device according to claim 6 is it is characterised in that described first cluster module includes:
Determining module, for determining target area to be split in described original target image according to the described geometric properties extracting
And nontarget area；
Filling module, is filled with for the color that described target area is different with described nontarget area brightness value with area
Divide described target area and described nontarget area；
Fuzzy Processing module, for the image after filling is carried out Fuzzy Processing, calculates every in the image after carrying out Fuzzy Processing
The corresponding brightness value of individual pixel；
Second cluster module, for the color characteristic according to each pixel described and corresponding described brightness value to described original
Each pixel in target image is clustered.
8. device according to claim 7 is it is characterised in that described second cluster module includes:
According to the corresponding brightness value of each pixel described, initial degree of membership determining module, for determining that each pixel is subordinate to respectively
Belong to the initial degree of membership of described target area and described nontarget area；
Initial centered value determining module, for the color characteristic according to each pixel described and corresponding described brightness value difference
Determine the initial centered value of target area and nontarget area；
3rd cluster module, for the face according to described initial degree of membership and described initial centered value and each pixel described
Color characteristic and corresponding described brightness value carry out cluster and obtain each pixel being under the jurisdiction of to the pixel in original target image
The probability of described target area.
9. device according to claim 7 is it is characterised in that described Fuzzy Processing module includes:
Blur radius acquisition module, for obtaining default blur radius, enters to the image after filling according to described blur radius
Row Fuzzy Processing；
Brightness value computing module, carries out the corresponding brightness value of each pixel in the image after described Fuzzy Processing for calculating.
10. device according to claim 6 it is characterised in that described segmentation module be additionally operable to according to cluster obtain every
The probability that individual pixel is under the jurisdiction of described target area carries out the segmentation in region, wherein, will be greater than the pixel of predetermined probabilities value
It is divided into target area, the pixel less than or equal to predetermined probabilities value is divided into nontarget area.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610702051.8A CN106340023B (en) | 2016-08-22 | 2016-08-22 | The method and apparatus of image segmentation |
| PCT/CN2017/098417 WO2018036462A1 (en) | 2016-08-22 | 2017-08-22 | Image segmentation method, computer apparatus, and computer storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610702051.8A CN106340023B (en) | 2016-08-22 | 2016-08-22 | The method and apparatus of image segmentation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106340023A true CN106340023A (en) | 2017-01-18 |
| CN106340023B CN106340023B (en) | 2019-03-05 |
Family
ID=57824596
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610702051.8A Active CN106340023B (en) | 2016-08-22 | 2016-08-22 | The method and apparatus of image segmentation |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN106340023B (en) |
| WO (1) | WO2018036462A1 (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018036462A1 (en) * | 2016-08-22 | 2018-03-01 | 腾讯科技（深圳）有限公司 | Image segmentation method, computer apparatus, and computer storage medium |
| CN110290426A (en) * | 2019-06-24 | 2019-09-27 | 腾讯科技（深圳）有限公司 | Method, apparatus, equipment and the storage medium of showing resource |
| CN110852938A (en) * | 2019-10-28 | 2020-02-28 | 腾讯科技（深圳）有限公司 | Display picture generation method and device and storage medium |
| CN110910400A (en) * | 2019-10-29 | 2020-03-24 | 北京三快在线科技有限公司 | Image processing method, image processing device, storage medium and electronic equipment |
| CN111079637A (en) * | 2019-12-12 | 2020-04-28 | 武汉轻工大学 | Method, device and equipment for segmenting rape flowers in field image and storage medium |
| CN111667553A (en) * | 2020-06-08 | 2020-09-15 | 北京有竹居网络技术有限公司 | Head-pixelized face color filling method and device and electronic equipment |
| CN112991357A (en) * | 2019-12-18 | 2021-06-18 | 中国船舶重工集团公司第七一一研究所 | Image segmentation method, system, computer device, readable storage medium and ship |
| WO2021253373A1 (en) * | 2020-06-19 | 2021-12-23 | Alibaba Group Holding Limited | Probabilistic geometric partitioning in video coding |
| WO2023108444A1 (en) * | 2021-12-14 | 2023-06-22 | 深圳传音控股股份有限公司 | Image processing method, intelligent terminal, and storage medium |
| CN116563312A (en) * | 2023-07-11 | 2023-08-08 | 山东古天电子科技有限公司 | Method for dividing display image of double-screen machine |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109147011B (en) * | 2018-08-27 | 2023-11-14 | 平安科技（深圳）有限公司 | License plate image generation method, license plate image generation device, computer equipment and storage medium |
| CN112308938B (en) * | 2019-07-30 | 2024-06-04 | 西安诺瓦星云科技股份有限公司 | Image processing method and image processing apparatus |
| CN111026641B (en) * | 2019-11-14 | 2023-06-20 | 北京云聚智慧科技有限公司 | Picture comparison method and electronic equipment |
| CN111325691B (en) * | 2020-02-20 | 2023-11-10 | Oppo广东移动通信有限公司 | Image correction method, apparatus, electronic device, and computer-readable storage medium |
| CN113012188A (en) * | 2021-03-23 | 2021-06-22 | 影石创新科技股份有限公司 | Image fusion method and device, computer equipment and storage medium |
| CN113344961B (en) * | 2021-06-01 | 2023-09-26 | 中国平安人寿保险股份有限公司 | Image background segmentation method, device, computing equipment and storage medium |
| CN114298937B (en) * | 2021-12-29 | 2024-05-17 | 深圳软牛科技集团股份有限公司 | JPEG photo restoration method and device and related components |
| CN114495236B (en) * | 2022-02-11 | 2023-02-28 | 北京百度网讯科技有限公司 | Image segmentation method, apparatus, device, medium, and program product |
| CN115439846B (en) * | 2022-08-09 | 2023-04-25 | 北京邮电大学 | Image segmentation method and device, electronic equipment and medium |
| CN116152562B (en) * | 2023-02-23 | 2023-08-01 | 北京朗视仪器股份有限公司 | Method and system for rapid color classification of color structured light image |
| CN116563326A (en) * | 2023-04-28 | 2023-08-08 | 江西绿萌科技控股有限公司 | Multi-target image segmentation method for fruit and vegetable sorting equipment |
| CN117765008B (en) * | 2023-12-26 | 2024-09-06 | 西北核技术研究所 | Image segmentation method based on self-adaptive feature perception |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020975A (en) * | 2012-12-29 | 2013-04-03 | 北方工业大学 | Wharf and ship segmentation method combining multi-source remote sensing image characteristics |
| CN103034872A (en) * | 2012-12-26 | 2013-04-10 | 四川农业大学 | Farmland pest recognition method based on colors and fuzzy clustering algorithm |
| CN103903257A (en) * | 2014-02-27 | 2014-07-02 | 西安电子科技大学 | Image segmentation method based on geometric block spacing symbiotic characteristics and semantic information |
| US20160155010A1 (en) * | 2014-11-28 | 2016-06-02 | Canon Kabushiki Kaisha | Image feature extraction method, image feature extraction apparatus, and recording medium storing program therefor |
| CN105869175A (en) * | 2016-04-21 | 2016-08-17 | 北京邮电大学 | Image segmentation method and system |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8539357B2 (en) * | 2007-11-21 | 2013-09-17 | Qualcomm Incorporated | Media preferences |
| CN202110564U (en) * | 2011-06-24 | 2012-01-11 | 华南理工大学 | Intelligent household voice control system combined with video channel |
| CN104134080B (en) * | 2014-08-01 | 2018-09-11 | 重庆大学 | A kind of road foundation collapses automatic testing method and system with slope failure |
| CN105469356B (en) * | 2015-11-23 | 2018-12-18 | 小米科技有限责任公司 | Face image processing process and device |
| CN106340023B (en) * | 2016-08-22 | 2019-03-05 | 腾讯科技（深圳）有限公司 | The method and apparatus of image segmentation |
-
2016
- 2016-08-22 CN CN201610702051.8A patent/CN106340023B/en active Active
-
2017
- 2017-08-22 WO PCT/CN2017/098417 patent/WO2018036462A1/en active Application Filing
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103034872A (en) * | 2012-12-26 | 2013-04-10 | 四川农业大学 | Farmland pest recognition method based on colors and fuzzy clustering algorithm |
| CN103020975A (en) * | 2012-12-29 | 2013-04-03 | 北方工业大学 | Wharf and ship segmentation method combining multi-source remote sensing image characteristics |
| CN103903257A (en) * | 2014-02-27 | 2014-07-02 | 西安电子科技大学 | Image segmentation method based on geometric block spacing symbiotic characteristics and semantic information |
| US20160155010A1 (en) * | 2014-11-28 | 2016-06-02 | Canon Kabushiki Kaisha | Image feature extraction method, image feature extraction apparatus, and recording medium storing program therefor |
| CN105869175A (en) * | 2016-04-21 | 2016-08-17 | 北京邮电大学 | Image segmentation method and system |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018036462A1 (en) * | 2016-08-22 | 2018-03-01 | 腾讯科技（深圳）有限公司 | Image segmentation method, computer apparatus, and computer storage medium |
| CN110290426B (en) * | 2019-06-24 | 2022-04-19 | 腾讯科技（深圳）有限公司 | Method, device and equipment for displaying resources and storage medium |
| CN110290426A (en) * | 2019-06-24 | 2019-09-27 | 腾讯科技（深圳）有限公司 | Method, apparatus, equipment and the storage medium of showing resource |
| CN110852938A (en) * | 2019-10-28 | 2020-02-28 | 腾讯科技（深圳）有限公司 | Display picture generation method and device and storage medium |
| CN110852938B (en) * | 2019-10-28 | 2024-03-19 | 腾讯科技（深圳）有限公司 | Display picture generation method, device and storage medium |
| CN110910400A (en) * | 2019-10-29 | 2020-03-24 | 北京三快在线科技有限公司 | Image processing method, image processing device, storage medium and electronic equipment |
| CN111079637A (en) * | 2019-12-12 | 2020-04-28 | 武汉轻工大学 | Method, device and equipment for segmenting rape flowers in field image and storage medium |
| CN111079637B (en) * | 2019-12-12 | 2023-09-08 | 武汉轻工大学 | Method, device, equipment and storage medium for segmenting rape flowers in field image |
| CN112991357A (en) * | 2019-12-18 | 2021-06-18 | 中国船舶重工集团公司第七一一研究所 | Image segmentation method, system, computer device, readable storage medium and ship |
| CN112991357B (en) * | 2019-12-18 | 2023-04-18 | 中国船舶集团有限公司第七一一研究所 | Image segmentation method, system, computer device, readable storage medium and ship |
| CN111667553A (en) * | 2020-06-08 | 2020-09-15 | 北京有竹居网络技术有限公司 | Head-pixelized face color filling method and device and electronic equipment |
| WO2021253373A1 (en) * | 2020-06-19 | 2021-12-23 | Alibaba Group Holding Limited | Probabilistic geometric partitioning in video coding |
| WO2023108444A1 (en) * | 2021-12-14 | 2023-06-22 | 深圳传音控股股份有限公司 | Image processing method, intelligent terminal, and storage medium |
| CN116563312A (en) * | 2023-07-11 | 2023-08-08 | 山东古天电子科技有限公司 | Method for dividing display image of double-screen machine |
| CN116563312B (en) * | 2023-07-11 | 2023-09-12 | 山东古天电子科技有限公司 | Method for dividing display image of double-screen machine |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106340023B (en) | 2019-03-05 |
| WO2018036462A1 (en) | 2018-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106340023B (en) | The method and apparatus of image segmentation | |
| CN105631880B (en) | Lane line dividing method and device | |
| CN103914699B (en) | A kind of method of the image enhaucament of the automatic lip gloss based on color space | |
| US9483835B2 (en) | Depth value restoration method and system | |
| CN104966085B (en) | A kind of remote sensing images region of interest area detecting method based on the fusion of more notable features | |
| CN108537239B (en) | Method for detecting image saliency target | |
| CN107392968B (en) | The image significance detection method of Fusion of Color comparison diagram and Color-spatial distribution figure | |
| CN105335719A (en) | Living body detection method and device | |
| CN104331690B (en) | A kind of colour of skin method for detecting human face and system based on single image | |
| Shih et al. | Automatic reference color selection for adaptive mathematical morphology and application in image segmentation | |
| CN112906550B (en) | Static gesture recognition method based on watershed transformation | |
| CN106067016B (en) | A kind of facial image eyeglass detection method and device | |
| CN109920018A (en) | Black-and-white photograph color recovery method, device and storage medium neural network based | |
| CN107895140A (en) | Porny identification method based on face complexion | |
| CN111160194A (en) | Static gesture image recognition method based on multi-feature fusion | |
| Huo et al. | Semisupervised learning based on a novel iterative optimization model for saliency detection | |
| CN111339932B (en) | Palm print image preprocessing method and system | |
| CN105488475A (en) | Method for detecting human face in mobile phone | |
| Raval et al. | Color image segmentation using FCM clustering technique in RGB, L* a* b, HSV, YIQ color spaces | |
| CN114359323A (en) | Image target area detection method based on visual attention mechanism | |
| CN106980371A (en) | It is a kind of based on the mobile augmented reality exchange method for closing on heterogeneous distributed structure | |
| Yang et al. | EHNQ: Subjective and objective quality evaluation of enhanced night-time images | |
| Manaf et al. | Color recognition system with augmented reality concept and finger interaction: Case study for color blind aid system | |
| Li et al. | A Student's t‐based density peaks clustering with superpixel segmentation (tDPCSS) method for image color clustering | |
| CN114359244B (en) | Image saliency detection method based on super-pixel segmentation and multiple color features |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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: 20210924 Address after: 518057 Tencent Building, No. 1 High-tech Zone, Nanshan District, Shenzhen City, Guangdong Province, 35 floors Patentee after: TENCENT TECHNOLOGY (SHENZHEN) Co.,Ltd. Patentee after: TENCENT CLOUD COMPUTING (BEIJING) Co.,Ltd. Address before: 2, 518000, East 403 room, SEG science and Technology Park, Zhenxing Road, Shenzhen, Guangdong, Futian District Patentee before: TENCENT TECHNOLOGY (SHENZHEN) Co.,Ltd. |