CN103198470A - Image cutting method and image cutting system - Google Patents
Image cutting method and image cutting system Download PDFInfo
- Publication number
- CN103198470A CN103198470A CN2013100607039A CN201310060703A CN103198470A CN 103198470 A CN103198470 A CN 103198470A CN 2013100607039 A CN2013100607039 A CN 2013100607039A CN 201310060703 A CN201310060703 A CN 201310060703A CN 103198470 A CN103198470 A CN 103198470A
- Authority
- CN
- China
- Prior art keywords
- point
- image
- limit
- pixel
- split
- 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
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000005520 cutting process Methods 0.000 title abstract description 7
- 238000012545 processing Methods 0.000 claims abstract description 7
- 238000010586 diagram Methods 0.000 abstract description 8
- 238000010276 construction Methods 0.000 abstract 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 26
- 230000005540 biological transmission Effects 0.000 description 10
- 238000000638 solvent extraction Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000012217 deletion Methods 0.000 description 2
- 230000037430 deletion Effects 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012271 agricultural production Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
Abstract
The invention provides an image cutting method and an image cutting system. The method includes the following steps of providing an image to be cut, obtaining any foreground point and a background point from a pixel point of the image to be cut, calibrating the foreground point and the background point, constructing a flow network diagram according to the pixel point of the image to be cut, the calibrated foreground point and the calibrated background point, processing the flow network diagram to obtain the minimum cut set by means of a push-relabel method, wherein the minimum cut set is a set of edges which have the minimum similarity and need to be cut in the flow network diagram when the image to be cut is cut into a foreground and a background; and cutting the image to be a foreground area and a background area according to the minimum cut set. By means of the method and construction of the flow network diagram of the image to be recognized, the push-relabel method is carried out, complexity of an algorithm is simplified, and meanwhile the method is suitable for all images and is convenient to use for users.
Description
Technical field
The present invention relates to the image Segmentation Technology field, particularly a kind of dividing method of image and system.
Background technology
Digital picture refers to the image represented with the two-dimensional array form.Digital picture can be generated by many different input equipments and technology, for example, and digital camera, scanner, coordinate measuring machine etc.Originally utilize computing machine to come processing graphics and image information.Nowadays, Digital Image Processing has a wide application in that national defence, industrial and agricultural production, life ﹠ amusement etc. are multi-field.
Image is cut apart the process that digital picture is subdivided into the set of a plurality of image region pixels that refers to.Its objective is the representation of simplifying or changing image, make image be more readily understood and analyze.Image cuts apart that in practice application comprises medical image, positioning object, recognition of face, traffic control system etc. in satellite image.
To the research of the image segmentation algorithm history of existing decades, by various theories thousands of kinds of various types of partitioning algorithms have been proposed.But these methods therefore can't be common to all images, and the complexity of conventional images partitioning algorithm have higher mostly at particular problem.
Summary of the invention
Purpose of the present invention is intended to solve at least one of above-mentioned technological deficiency.
For achieving the above object, the embodiment of one aspect of the present invention proposes a kind of dividing method of image, may further comprise the steps: image to be split is provided; From the pixel of described image to be split, obtain any foreground point and background dot, and described foreground point and background dot are demarcated; According to the pixel of described image to be split and described foreground point and the background dot structure flow network figure of demarcation; Employing is pressed into heavy mask method described flow network figure is handled to obtain minimal cut set, and wherein, minimal cut set is for being divided into described image to be split the set on prospect and background limit of the similarity minimum of required disconnection in described flow network figure; And according to described minimal cut set described image to be split is divided into foreground area and background area.
According to the method for the embodiment of the invention, by constructing the flow network figure of image to be identified, and it is pressed into and heavy mark, therefore simplified the complexity of algorithm, be applicable to that simultaneously all images are convenient for users.
In an example of the present invention, in described flow network figure, the weight on the limit between the pixel that each pixel is adjacent is W (p, q), and except the weight on other pixels of the consecutive point of source point and meeting point and the limit between the described source point and to remove the consecutive point of meeting point and other pixels of source point and the weight on the limit between the described meeting point be constant, wherein, described source point is described foreground point, and described meeting point is sight spot, described back.
In an example of the present invention, the weights W between the described neighbor pixel (p q) represents by following formula,
Wherein, (σ represents to regulate parameter, I to dist for p, the q) distance between expression p and the q
pThe brightness that expression p is ordered, I
qThe brightness of the point of expression q.
In an example of the present invention, described constant represents by following formula,
Wherein, the adjacent point set that N (p) expression p is ordered, V represents the set of pixel, W (p, q) expression point p, the weight on limit between the q.
For achieving the above object, embodiments of the invention propose a kind of segmenting system of image on the other hand, comprising: acquisition module is used for providing image to be split; Demarcating module is used for obtaining any foreground point and background dot from the pixel of described image to be split, and described foreground point and background dot is demarcated; Constructing module is used for according to the pixel of described image to be split and described foreground point and the background dot structure flow network figure of demarcation; Processing module, be pressed into heavy mask method for employing described flow network figure is handled to obtain minimal cut set, wherein, minimal cut set is for being divided into described image to be split the set on prospect and background limit of the similarity minimum of required disconnection in described flow network figure; And cut apart module, be used for according to described minimal cut set described image to be split being divided into foreground area and background area.
In an example of the present invention, in described flow network figure, the weight on the limit between the pixel that each pixel is adjacent is W (p, q), and except the weight on other pixels of the consecutive point of source point and meeting point and the limit between the described source point and to remove the consecutive point of meeting point and other pixels of source point and the weight on the limit between the described meeting point be constant, wherein, described source point is described foreground point, and described meeting point is sight spot, described back.
In an example of the present invention, the weights W between the described neighbor pixel (p q) represents by following formula,
Wherein, (σ represents to regulate parameter, I to sist for p, the q) distance between expression p and the q
pThe brightness that expression p is ordered, I
qThe brightness of the point of expression q.
In an example of the present invention, described constant represents by following formula,
Wherein, the adjacent point set that N (p) expression p is ordered, V represents the set of pixel, W (p, q) expression point p, the weight on limit between the q.
According to the system of the embodiment of the invention, by constructing the flow network figure of image to be identified, and it is pressed into and heavy mark, therefore simplified the complexity of algorithm, be applicable to that simultaneously all images are convenient for users.
The aspect that the present invention adds and advantage part in the following description provide, and part will become obviously from the following description, or recognize by practice of the present invention.
Description of drawings
Above-mentioned and/or the additional aspect of the present invention and advantage are from obviously and easily understanding becoming the description of embodiment below in conjunction with accompanying drawing, wherein:
Fig. 1 is the process flow diagram of the dividing method of image according to an embodiment of the invention;
Fig. 2 is that synoptic diagram is divided on the limit of image to be split according to an embodiment of the invention;
Fig. 3 is the synoptic diagram of cutting apart of image to be split according to an embodiment of the invention; And
Fig. 4 is the structured flowchart of the segmenting system of image according to an embodiment of the invention.
Embodiment
Describe embodiments of the invention below in detail, the example of embodiment is shown in the drawings, and wherein identical or similar label is represented identical or similar elements or the element with identical or similar functions from start to finish.Be exemplary below by the embodiment that is described with reference to the drawings, only be used for explaining the present invention, and can not be interpreted as limitation of the present invention.
Fig. 1 is the process flow diagram of similar image sorting technique according to an embodiment of the invention.As shown in Figure 1, the similar image sorting technique according to the embodiment of the invention may further comprise the steps:
Step S101 provides image to be split.
Step S102 obtains any foreground point and background dot from the pixel of image to be split, and foreground point and background dot are demarcated.
Particularly, in image to be split, determine a foreground point and a background dot arbitrarily.Can adopt interactive mode, or priori, for example extract a point respectively out as a setting and the foreground point in four angles and the central area.
Step S103 is according to the pixel of image to be split and foreground point and the background dot structure flow network figure of demarcation.
Particularly, the set that comprises all pixels in the image to be split is vertex set V, and is source point with foreground point S, and background dot T is meeting point structure flow network figure G=＜V, E 〉.This flow network figure G is undirected weighted graph, and the weight of limit E is represented the maximum flow that this limit can be passed through.
The limit collection E of flow network figure G constitutes, there is a limit between each pixel and its each adjacent pixels point, wherein adjacent definition can be 4 neighborhoods (up and down) or 8 neighborhoods (that 4 neighborhoods add is upper left, lower-left, upper right, bottom right), for example, in Fig. 2, represent adjacent edge with fine line.Adjacent 2 p in the image to be split, the limit weight between q be W (p q) can represent by following formula,
Wherein, (σ represents to regulate parameter, I to dist for p, the q) distance between expression p and the q
pThe brightness that expression p is ordered, I
qThe brightness of the point of expression q.This apart from dist (p, q) can pass through the Euclidean distance formula
City range formula dist (p, q)=| x
p-x
q|+y
p-y
q|, or chessboard distance formula dist (p, q)=max (| x
p-x
q|, | y
p-y
q|) in any one draw, wherein, (x
p, y
p) and (x
q, y
q) represent some p respectively and put the coordinate of q in image.
Except the weight on other pixels of the consecutive point of source point S and meeting point T and the limit (representing with heavy line among Fig. 2) between the source point S and to remove the consecutive point of meeting point T and other pixels of source point S and the weight on the limit (dotting among Fig. 2) between the meeting point T be that constant K can be represented by following formula
Wherein, (σ represents to regulate parameter, I to dist for p, the q) distance between expression p and the q
pThe brightness that expression p is ordered, I
qThe brightness of the point of expression q.In the implication of this formula for to ask the limit weight sum that is adjacent pixel for a pixel, all pixels in the traversing graph picture, should and maximal value add 1 and be K.
Step S104, employing is pressed into heavy mask method convection current network chart and handles to obtain minimal cut set, and wherein, minimal cut set is for being divided into image to be split the set on prospect and background limit of the similarity minimum of required disconnection in flow network figure.
Particularly, in structure flow network figure, the weight size on every limit shows the similarity degree of two points that this limit connects.Minimal cut set between background dot and foreground point is representing the set on the limit of the similarity minimum that the prospect that flow network figure can be divided into and background two parts need disconnect.Try to achieve the minimal cut of this flow network, just equal to have tried to achieve split image.According to the fundamental theorem of flow network graph theory, the flow sum of minimal cut equals the max-flow in the network.When flow network reaches max-flow, all limits that have cut set state that all reaches capacity, this cut set is exactly minimal cut.Therefore finding the solution minimal cut is converted into and finds the solution max-flow.
Be pressed into heavily to mark the summit is analogized to reservoir, the limit is analogized to pipeline between reservoir, the weight on limit is analogized to the maximum flow that pipeline can pass through.Current flow out from source point, through Internet communication, finally flow into meeting point.Each reservoir has certain altitude, and water can only be toward lower stream.The water cut of each reservoir is the poor of inflow reservoir and the flow that flows out reservoir.
P, q represent the point among the flow network figure G: the height d (p) of each point; The water cut e of each point (p); The flow F of point-to-point transmission (p, q).In the solution procedure, the flow on the flow network must satisfy the restrictive condition that aforesaid flow network figure G produces all the time.Be example with Fig. 2, among (1) flow network figure G 2 when having the limit directly to link to each other, point-to-point transmission just allows to exist flow, for example allows to exist flow between ab, does not allow to exist flow between ac.(2) flow of point-to-point transmission must not surpass the weight on point-to-point transmission limit, F(a for example, b)＜=W(a, b), F(S, a)＜=K, wherein, F (p, the q) flow between expression point p and q, the height of d (p) expression point p, the water cut of e (p) expression point p.
At first, the height of source point S is made as the value (N is the number of flow network figure G mid point) that is not less than N-1, the height of all the other points is made as 0, and the water cut of each point is made as 0, and the flow of point-to-point transmission is made as 0 arbitrarily.Then, each the some p for linking to each other with source point S is made as the upper limit of the two flow with the water cut of this point, i.e. the weight on limit, e (p)=W (S p), and is made as the upper limit of the two flow with the two flow, namely F (S, p)=W (S, p).Flow network figure mid point is pressed into or heavily marks.For a p and some q, be that the p point is not source point S or meeting point T from the condition that a p is pressed into a q with water; The p point still has water, i.e. e (p)〉0; Pipeline between p and q still has surplus, namely W (p, q)-F (p, q)〉0; (4) the q aspect ratio p of the ordering height of ordering is little by 1, i.e. d (p)-d (q)=1.
For a p and some q, be at first to get the smaller value of the water yield that can also flow through between p point water cut and pq from the step that a p is pressed into a q with water, be designated as f, be f=min (e (p), W (p, q)-F (p, q)) thus the p water cut of ordering reduce f, be e (p)=e (p)-f, and the flow between pq increase f, i.e. F (p, q)=and F (p, q)+f.
In one embodiment of the invention, the condition that a p is heavily marked is that (1) p point is not source point S or meeting point T; (2) the p point still has water, i.e. e (p)〉0; (3) can't carry out push operation from the p point to coupled point.The step that a p is heavily marked is, at first find out link to each other with p and and p between pipeline also have the point set of surplus, use R(p) expression, namely
Then, change the height of p into R(p) in the height of minimum point add 1, namely
Be pressed into or heavy mask method repeats operation by above, when satisfy without any point be pressed into or when heavily marking condition the flow on the gained flow network figure be maximum flow.
Among the flow network figure at this moment, searching can be cut off flow network to this and not comprise source point S and the two-part minimal cut set of meeting point T, and corresponding in the true picture is exactly the partitioning boundary place.Particularly, in current flow network figure, find all flows to reach the limit of the upper limit, namely satisfy F (p, q)=W (p, all limits q), and this limit progressively added minimal cut set by flow order from small to large, up to flow network figure being broken as the two parts that comprise source point S and meeting point T respectively.Afterwards again, flow network figure is broken as under the two-part situation guaranteeing by flow order from big to small, progressively with the redundant limit deletion in the minimal cut set, finally obtains minimal cut set.
Step S105 is divided into foreground area and background area according to minimal cut set with image to be split.
As shown in Figure 3, when the minimal cut set that finds comprise (S, d), (S, g), (S, h), (S, e), (S, f), (S, i), (T, a), (T, b), (T, c), (a, d), (b, e), (c, f) 12 limits then correspond in the image to be split, should cut apart in the thickest horizontal line position as shown in Figure 3, S, a, b, c constitute foreground area, and all the other points constitute the background area.
According to the method for the embodiment of the invention, by constructing the flow network figure of image to be identified, and it is pressed into and heavy mark, therefore simplified the complexity of algorithm, be applicable to that simultaneously all images are convenient for users.
Fig. 4 is the structured flowchart of the segmenting system of image according to an embodiment of the invention.As shown in Figure 4, the segmenting system according to the image of the embodiment of the invention comprises acquisition module 100, demarcating module 200, constructing module 300, processing module 400 and cuts apart module 500.
Acquisition module 100 is used for providing image to be split.
Demarcating module 200 is used for obtaining any foreground point and background dot from the pixel of image to be split, and foreground point and background dot are demarcated.
Particularly, in image to be split, determine a foreground point and a background dot arbitrarily.Can adopt interactive mode, or priori, for example extract a point respectively out as a setting and the foreground point in four angles and the central area.
Constructing module 300 is used for according to the foreground point of the pixel of image to be split and demarcation and background dot structure flow network figure.
Particularly, the set that comprises all pixels in the image to be split is vertex set V, and is source point with foreground point S, and background dot T is meeting point structure flow network figure G=＜V, E 〉.This flow network figure G is undirected weighted graph, and the weight of limit E is represented the maximum flow that this limit can be passed through.
The limit collection E of flow network figure G constitutes, there is a limit between each pixel and its each adjacent pixels point, wherein adjacent definition can be 4 neighborhoods (up and down) or 8 neighborhoods (that 4 neighborhoods add is upper left, lower-left, upper right, bottom right), for example, in Fig. 2, represent adjacent edge with fine line.Adjacent 2 p in the image to be split, the limit weight between q be W (p q) can represent by following formula,
Wherein, (σ represents to regulate parameter, I to dist for p, the q) distance between expression p and the q
pThe brightness that expression p is ordered, I
qThe brightness of the point of expression q.This apart from dist (p, q) can pass through the Euclidean distance formula
City range formula dist (p, q)=| x
p-x
q|+y
p-y
q|, or chessboard distance formula dist (p, q)=max (| x
p-x
q|, | y
p-y
q|) in any one draw, wherein, (x
p, y
p) and (x
q, y
q) represent some p respectively and put the coordinate of q in image.
Except the weight on other pixels of the consecutive point of source point S and meeting point T and the limit (representing with heavy line among Fig. 2) between the source point S and to remove the consecutive point of meeting point T and other pixels of source point S and the weight on the limit (dotting among Fig. 2) between the meeting point T be that constant K can be represented by following formula
Wherein, (σ represents to regulate parameter, I to dist for p, the q) distance between expression p and the q
pThe brightness that expression p is ordered, I
qThe brightness of the point of expression q.In the implication of this formula for to ask the limit weight sum that is adjacent pixel for a pixel, all pixels in the traversing graph picture, should and maximal value add 1 and be K.
Processing module 400 is pressed into heavy mask method convection current network chart for employing and handles to obtain minimal cut set, and wherein, minimal cut set is for being divided into image to be split the set on prospect and background limit of the similarity minimum of required disconnection in flow network figure.
Particularly, in structure flow network figure, the weight size on every limit shows the similarity degree of two points that this limit connects.Minimal cut set between background dot and foreground point is representing the set on the limit of the similarity minimum that the prospect that flow network figure can be divided into and background two parts need disconnect.Try to achieve the minimal cut of this flow network, just equal to have tried to achieve split image.According to the fundamental theorem of flow network graph theory, the flow sum of minimal cut equals the max-flow in the network.When flow network reaches max-flow, all limits that have cut set state that all reaches capacity, this cut set is exactly minimal cut.Therefore finding the solution minimal cut is converted into and finds the solution max-flow.
Be pressed into heavily to mark the summit is analogized to reservoir, the limit is analogized to pipeline between reservoir, the weight on limit is analogized to the maximum flow that pipeline can pass through.Current flow out from source point, through Internet communication, finally flow into meeting point.Each reservoir has certain altitude, and water can only be toward lower stream.The water cut of each reservoir is the poor of inflow reservoir and the flow that flows out reservoir.
P, q represent the point among the flow network figure G: the height d (p) of each point; The water cut e of each point (p); The flow F of point-to-point transmission (p, q).In the solution procedure, the flow on the flow network must satisfy the restrictive condition that aforesaid flow network figure G produces all the time.Be example with Fig. 2, among (1) flow network figure G 2 when having the limit directly to link to each other, point-to-point transmission just allows to exist flow, for example allows to exist flow between ab, does not allow to exist flow between ac.(2) flow of point-to-point transmission must not surpass the weight on point-to-point transmission limit, F(a for example, b)＜=W(a, b), F(S, a)＜=K, wherein, F (p, the q) flow between expression point p and q, the height of d (p) expression point p, the water cut of e (p) expression point p.
At first, the height of source point S is made as the value (N is the number of flow network figure G mid point) that is not less than N-1, the height of all the other points is made as 0, and the water cut of each point is made as 0, and the flow of point-to-point transmission is made as 0 arbitrarily.Then, each the some p for linking to each other with source point S is made as the upper limit of the two flow with the water cut of this point, i.e. the weight on limit, e (p)=W (S p), and is made as the upper limit of the two flow with the two flow, namely F (S, p)=W (S, p).Flow network figure mid point is pressed into or heavily marks.For a p and some q, be that the p point is not source point S or meeting point T from the condition that a p is pressed into a q with water; The p point still has water, i.e. e (p)〉0; Pipeline between p and q still has surplus, namely W (p, q)-F (p, q)〉0; (4) the q aspect ratio p of the ordering height of ordering is little by 1, i.e. d (p)-d (q)=1.
For a p and some q, be at first to get the smaller value of the water yield that can also flow through between p point water cut and pq from the step that a p is pressed into a q with water, be designated as f, be f=min (e (p), W (p, q)-F (p, q)) thus the p water cut of ordering reduce f, be e (p)=e (p)-f, and the flow between pq increase f, i.e. F (p, q)=and F (p, q)+f.
In one embodiment of the invention, the condition that a p is heavily marked is that (1) p point is not source point S or meeting point T; (2) the p point still has water, i.e. e (p)〉0; (3) can't carry out push operation from the p point to coupled point.The step that a p is heavily marked is, at first find out link to each other with p and and p between pipeline also have the point set of surplus, use R(p) expression, namely
Then, change the height of p into R(p) in the height of minimum point add 1, namely
Be pressed into or heavy mask method repeats operation by above, when satisfy without any point be pressed into or when heavily marking condition the flow on the gained flow network figure be maximum flow.
Among the flow network figure at this moment, searching can be cut off flow network to this and not comprise source point S and the two-part minimal cut set of meeting point T, and corresponding in the true picture is exactly the partitioning boundary place.Particularly, in current flow network figure, find all flows to reach the limit of the upper limit, namely satisfy F (p, q)=W (p, all limits q), and this limit progressively added minimal cut set by flow order from small to large, up to flow network figure being broken as the two parts that comprise source point S and meeting point T respectively.Afterwards again, flow network figure is broken as under the two-part situation guaranteeing by flow order from big to small, progressively with the redundant limit deletion in the minimal cut set, finally obtains minimal cut set.
Cutting apart module 500 is used for according to minimal cut set image to be split being divided into foreground area and background area.
As shown in Figure 3, when the minimal cut set that finds comprise (S, d), (S, g), (S, h), (S, e), (S, f), (S, i), (T, a), (T, b), (T, c), (a, d), (b, e), (c, f) 12 limits then correspond in the image to be split, should cut apart in the thickest horizontal line position as shown in Figure 3, S, a, b, c constitute foreground area, and all the other points constitute the background area.
According to the method for the embodiment of the invention, by constructing the flow network figure of image to be identified, and it is pressed into and heavy mark, therefore simplified the complexity of algorithm, be applicable to that simultaneously all images are convenient for users.
Although illustrated and described embodiments of the invention above, be understandable that, above-described embodiment is exemplary, can not be interpreted as limitation of the present invention, those of ordinary skill in the art can change above-described embodiment under the situation that does not break away from principle of the present invention and aim within the scope of the invention, modification, replacement and modification.
Claims (8)
1. the dividing method of an image is characterized in that, may further comprise the steps:
Image to be split is provided;
From the pixel of described image to be split, obtain any foreground point and background dot, and described foreground point and background dot are demarcated;
According to the pixel of described image to be split and described foreground point and the background dot structure flow network figure of demarcation;
Employing is pressed into heavy mask method described flow network figure is handled to obtain minimal cut set, and wherein, described minimal cut set is for being divided into described image to be split the set on prospect and background limit of the similarity minimum of required disconnection in described flow network figure; And
According to described minimal cut set described image to be split is divided into foreground area and background area.
2. the dividing method of image as claimed in claim 1, it is characterized in that, in described flow network figure, the weight on the limit between the pixel that each pixel is adjacent is W (p, q), and except the weight on other pixels of the consecutive point of source point and meeting point and the limit between the described source point and to remove the consecutive point of meeting point and other pixels of source point and the weight on the limit between the described meeting point be constant
Wherein, described source point is described foreground point, and described meeting point is sight spot, described back.
3. the dividing method of image as claimed in claim 1 is characterized in that, the weights W between the described neighbor pixel (p q) represents by following formula,
Wherein, (σ represents to regulate parameter, I to dist for p, the q) distance between expression p and the q
pThe brightness that expression p is ordered, I
qThe brightness of the point of expression q.
4. the dividing method of image as claimed in claim 1 is characterized in that, described constant represents by following formula,
Wherein, the adjacent point set that N (p) expression p is ordered, V represents the set of pixel, W (p, q) expression point p, the weight on limit between the q.
5. the segmenting system of an image is characterized in that, comprising:
Acquisition module is used for providing image to be split;
Demarcating module is used for obtaining any foreground point and background dot from the pixel of described image to be split, and described foreground point and background dot is demarcated;
Constructing module is used for according to the pixel of described image to be split and described foreground point and the background dot structure flow network figure of demarcation;
Processing module, be pressed into heavy mask method for employing described flow network figure is handled to obtain minimal cut set, wherein, described minimal cut set is for being divided into described image to be split the set on prospect and background limit of the similarity minimum of required disconnection in described flow network figure; And
Cut apart module, be used for according to described minimal cut set described image to be split being divided into foreground area and background area.
6. the segmenting system of image as claimed in claim 5, it is characterized in that, in described flow network figure, the weight on the limit between the pixel that each pixel is adjacent is W (p, q), and except the weight on other pixels of the consecutive point of source point and meeting point and the limit between the described source point and to remove the consecutive point of meeting point and other pixels of source point and the weight on the limit between the described meeting point be constant
Wherein, described source point is described foreground point, and described meeting point is sight spot, described back.
7. similar image categorizing system as claimed in claim 5 is characterized in that, the weights W between the described neighbor pixel (p q) represents by following formula,
Wherein, (σ represents to regulate parameter, I to dist for p, the q) distance between expression p and the q
pThe brightness that expression p is ordered, I
qThe brightness of the point of expression q.
8. similar image categorizing system as claimed in claim 5 is characterized in that, described constant represents by following formula,
Wherein, the adjacent point set that N (p) expression p is ordered, V represents the set of pixel, W (p, q) expression point p, the weight on limit between the q.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310060703.9A CN103198470B (en) | 2013-02-26 | 2013-02-26 | Image cutting method and image cutting system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310060703.9A CN103198470B (en) | 2013-02-26 | 2013-02-26 | Image cutting method and image cutting system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103198470A true CN103198470A (en) | 2013-07-10 |
| CN103198470B CN103198470B (en) | 2017-02-15 |
Family
ID=48720976
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310060703.9A Expired - Fee Related CN103198470B (en) | 2013-02-26 | 2013-02-26 | Image cutting method and image cutting system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103198470B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104134192A (en) * | 2014-07-23 | 2014-11-05 | 中国科学院深圳先进技术研究院 | Image defogging method and system |
| WO2015027412A1 (en) * | 2013-08-28 | 2015-03-05 | 华为技术有限公司 | Method and device for complex graph processing |
| CN106204600A (en) * | 2016-07-07 | 2016-12-07 | 广东技术师范学院 | Cerebral tumor image partition method based on multisequencing MR image related information |
| CN106920244A (en) * | 2017-01-13 | 2017-07-04 | 广州中医药大学 | A kind of method of background dot near detection image edges of regions |
| CN107204034A (en) * | 2016-03-17 | 2017-09-26 | 腾讯科技（深圳）有限公司 | A kind of image processing method and terminal |
| CN109740658A (en) * | 2018-12-28 | 2019-05-10 | 陕西师范大学 | A kind of semi-supervision image classification method based on weighted graph |
| CN111277902A (en) * | 2020-02-17 | 2020-06-12 | 北京达佳互联信息技术有限公司 | Video matching method, device and equipment |
| CN113139976A (en) * | 2021-04-20 | 2021-07-20 | 上海科技大学 | Maximum flow minimum cut solving algorithm capable of being terminated in advance based on push and re-marking |
| CN116258725A (en) * | 2023-05-16 | 2023-06-13 | 福建自贸试验区厦门片区Manteia数据科技有限公司 | Medical image processing method and device based on feature images and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2004003847A2 (en) * | 2002-06-28 | 2004-01-08 | Koninklijke Philips Electronics N.V. | Background-foreground segmentation based on object classification |
| CN101101668A (en) * | 2007-07-13 | 2008-01-09 | 杭州电子科技大学 | Medical image cutting method based on oscillatory network |
| US20100046830A1 (en) * | 2008-08-22 | 2010-02-25 | Jue Wang | Automatic Video Image Segmentation |
| CN102270338A (en) * | 2011-06-27 | 2011-12-07 | 西安交通大学 | Method for effectively segmenting repeated object based on image representation improvement |
| CN102467740A (en) * | 2010-11-08 | 2012-05-23 | 北京大学 | Foreground and background interactive segmentation method of large-size color image and system thereof |
-
2013
- 2013-02-26 CN CN201310060703.9A patent/CN103198470B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2004003847A2 (en) * | 2002-06-28 | 2004-01-08 | Koninklijke Philips Electronics N.V. | Background-foreground segmentation based on object classification |
| CN101101668A (en) * | 2007-07-13 | 2008-01-09 | 杭州电子科技大学 | Medical image cutting method based on oscillatory network |
| US20100046830A1 (en) * | 2008-08-22 | 2010-02-25 | Jue Wang | Automatic Video Image Segmentation |
| CN102467740A (en) * | 2010-11-08 | 2012-05-23 | 北京大学 | Foreground and background interactive segmentation method of large-size color image and system thereof |
| CN102270338A (en) * | 2011-06-27 | 2011-12-07 | 西安交通大学 | Method for effectively segmenting repeated object based on image representation improvement |
Non-Patent Citations (1)
| Title |
|---|
| 王建青: ""基于图割理论的图像分割方法研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015027412A1 (en) * | 2013-08-28 | 2015-03-05 | 华为技术有限公司 | Method and device for complex graph processing |
| CN104134192A (en) * | 2014-07-23 | 2014-11-05 | 中国科学院深圳先进技术研究院 | Image defogging method and system |
| CN107204034A (en) * | 2016-03-17 | 2017-09-26 | 腾讯科技（深圳）有限公司 | A kind of image processing method and terminal |
| CN107204034B (en) * | 2016-03-17 | 2019-09-13 | 腾讯科技（深圳）有限公司 | A kind of image processing method and terminal |
| US11037275B2 (en) | 2016-03-17 | 2021-06-15 | Tencent Technology (Shenzhen) Company Limited | Complex architecture for image processing |
| CN106204600A (en) * | 2016-07-07 | 2016-12-07 | 广东技术师范学院 | Cerebral tumor image partition method based on multisequencing MR image related information |
| CN106920244A (en) * | 2017-01-13 | 2017-07-04 | 广州中医药大学 | A kind of method of background dot near detection image edges of regions |
| CN106920244B (en) * | 2017-01-13 | 2019-08-02 | 广州中医药大学 | A kind of method of the neighbouring background dot of detection image edges of regions |
| CN109740658B (en) * | 2018-12-28 | 2023-04-18 | 陕西师范大学 | Semi-supervised image classification method based on weighted graph |
| CN109740658A (en) * | 2018-12-28 | 2019-05-10 | 陕西师范大学 | A kind of semi-supervision image classification method based on weighted graph |
| CN111277902A (en) * | 2020-02-17 | 2020-06-12 | 北京达佳互联信息技术有限公司 | Video matching method, device and equipment |
| CN111277902B (en) * | 2020-02-17 | 2022-03-25 | 北京达佳互联信息技术有限公司 | Video matching method, device and equipment |
| WO2022222367A1 (en) * | 2021-04-20 | 2022-10-27 | 上海科技大学 | Push-relabel-based max-flow min-cut solution algorithm capable of early termination |
| CN113139976A (en) * | 2021-04-20 | 2021-07-20 | 上海科技大学 | Maximum flow minimum cut solving algorithm capable of being terminated in advance based on push and re-marking |
| CN116258725A (en) * | 2023-05-16 | 2023-06-13 | 福建自贸试验区厦门片区Manteia数据科技有限公司 | Medical image processing method and device based on feature images and storage medium |
| CN116258725B (en) * | 2023-05-16 | 2023-08-22 | 福建自贸试验区厦门片区Manteia数据科技有限公司 | Medical image processing method and device based on feature images and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103198470B (en) | 2017-02-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103198470A (en) | Image cutting method and image cutting system | |
| CN109740465B (en) | Lane line detection algorithm based on example segmentation neural network framework | |
| CN105608456A (en) | Multi-directional text detection method based on full convolution network | |
| CN108090911B (en) | Near-shore ship segmentation method for optical remote sensing image | |
| CN105631880A (en) | Lane line segmentation method and apparatus | |
| CN102708370A (en) | Method and device for extracting multi-view angle image foreground target | |
| CN104408733B (en) | Object random walk-based visual saliency detection method and system for remote sensing image | |
| CN103985133A (en) | Search method and system for optimal splicing lines among images based on graph-cut energy optimization | |
| CN103389042A (en) | Ground automatic detecting and scene height calculating method based on depth image | |
| CN102609950A (en) | Two-dimensional video depth map generation process | |
| CN106097313B (en) | Image segmentation method and device | |
| CN105513105A (en) | Image background blurring method based on saliency map | |
| CN102542571A (en) | Moving target detecting method and device | |
| CN101510304A (en) | Method, device and pick-up head for dividing and obtaining foreground image | |
| CN102609723A (en) | Image classification based method and device for automatically segmenting videos | |
| CN102542544A (en) | Color matching method and system | |
| CN111028350A (en) | Method for constructing grid map by using binocular stereo camera | |
| CN105258681A (en) | Control for curve edge feature location and location method thereof | |
| CN104897671A (en) | Identification system for fruit stem and calyx of fruit | |
| CN106257537A (en) | A kind of spatial depth extracting method based on field information | |
| CN103578107A (en) | Method for interactive image segmentation | |
| CN116503836A (en) | 3D target detection method based on depth completion and image segmentation | |
| CN102982524A (en) | Splicing method for corn ear order images | |
| CN104637066A (en) | Method for extracting binary image quick skeleton based on sequential refining | |
| CN102136151B (en) | Method for vectorizing raster image |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170215 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |