Editing of presegemented images/volumes with the multilabel random walker or graph cut segmentations
View Patent ↗Presegmentation or prior, pre-existing segmentation of an object obtained through other means will be presented for interactive editing of a segmented object in an image or volume. Prior segmentation will be seamlessly combined with graph cuts or the random walker methods. Editing of the presegmentation is possible, while maintaining the important property of both methods that an arbitrary segmentation may be achieved with enough interaction.
1. A method for processing an object in a set of image data comprising:
presegmenting the object into a presegmentation using a first segmentation process;
analyzing the presegmentation and determining a weighting factor based on the analysis of the presegmentation;
placing one or more seeds in the set of image data;
segmenting by a processor of the object using a second segmentation process using the seeds and the weighting factor and the presegmentation, wherein the first and the second segmentation processes are different types of segmentation processes; and
wherein the weighting factor is a function of a distance of the presegmented object from one or more seeds, wherein the weighting factor is a function of a minimum distance between presegmentation and seeds expressed as
γ
i
=
κ
exp
(
-
d
(
v
i
,
v
j
)
σ
)
,
wherein the factors κ and σ are user controlled constants, γ i is the weighting factor v i and v j are nodes and d(v i , v j ) is the minimum distance from v i to all v j ε F, B with F and B being foreground and background respectively.
2. The method of claim 1 , wherein the weighting factor is interactively modified by changing κ or σ.
3. The method of claim 1 , wherein the second segmentation process is a graph cuts method.
4. The method of claim 1 , wherein the second segmentation process is a random walker method.
5. The method of claim 1 , wherein the step of segmenting the object using the second segmentation process with one or more seeds, the results of the first segmentation process and the weighting factor includes minimizing an energy expression
Q
(
x
)
=
∑
e
ij
w
ij
(
x
i
-
x
j
)
2
+
γ
(
∑
i
(
1
-
p
i
)
x
i
+
∑
i
p
i
(
1
-
x
i
)
)
,
wherein e ij denotes an edge, spanning two nodes, w ij is a weight of an edge e ij , p i denotes a presegmentation of a node, x i and x j represent a constraint for nodes v i and v i respectively and γ is the weighting factor.
6. The method of claim 5 , wherein the minimization of the energy expression is constrained by editing sets of nodes.
7. The method of claim 1 , wherein one or more additional seeds are added to the set of image data.
8. A system for segmenting an object in a set of image data using one or more presegmented objects, comprising:
a processor;
computer software operable on the processor, the computer software being capable of:
presegmenting the object into a presegmentation using a first segmentation process;
analyzing the presegmentation and determining a weighting factor based on the analysis of the presegmentation;
placing one or more seeds in the set of image data;
segmenting the object using a second segmentation process using the seeds and the weighting factor and the presegmentation, wherein the first and the second segmentation processes are different types of segmentation processes; and
wherein the step of segmenting the object using a second segmentation process with one or more seeds, the presegmentation and the weighting factor includes minimizing an energy expression
Q
(
x
)
=
∑
e
ij
w
ij
(
x
i
-
x
j
)
2
+
γ
(
∑
i
(
1
-
p
i
)
x
i
+
∑
i
p
i
(
1
-
x
i
)
)
,
wherein e ij denotes an edge spanning two nodes, w ij is a weight of an edge e denotes a presegmentation of a node, x i and x i represent a constraint for nodes v i and v i respectively and γ is the weighting factor.
9. The system of claim 8 , wherein the second segmentation process is a graph cuts method.
10. The system of claim 8 , wherein the second segmentation process is a random walker method.
11. The system of claim 8 , wherein the weighting factor is a function of a distance of the presegmented object from one or more seeds.
12. The system of claim 8 , wherein the weighting factor is a function of a minimum distance between presegmentation and seeds expressed as
y
i
=
κ
exp
(
-
d
(
v
i
,
v
j
)
σ
)
,
wherein the factors κ and σ are user controlled constants, γ i is the weighting factor v i and v j are nodes and d(v i ,v j ) is a minimum distance from v i to all v j ε F, B with F and B being foreground and background respectively.
13. The system of claim 12 , wherein the weighting factor is interactively modified by changing κ or σ.
14. The system of claim 8 , wherein the minimization of the energy expression is constrained by editing sets of nodes.
15. The system of claim 8 , wherein one or more seeds are added to the set of image data.
16. The system of claim 8 wherein the weighting factor is interactively modified.
17. A system for segmenting an object in a set of image data using one or more presegmented objects, comprising:
a processor;
computer software operable on the processor, the computer software being capable of:
presegmenting the object into a presegmentation by using a first segmentation process;
analyzing the presegmentation and determining a weighting factor based on the analysis of the presegmentation;
placing one or more seeds in the set of image data; and
segmenting the object using a second segmentation process using the seeds and the weighting factor and the results from the first segmentation process; and
wherein the weighting factor is a function of a minimum distance between presegmentation and seeds expressed as
γ
i
=
κ
exp
(
-
d
(
v
i
,
v
j
)
σ
)
,
and wherein the factors κ and σ are user controlled constants, γ i is the weighting factor v i and v j are nodes and d(v i ,v j ) is a minimum distance from v i to all v j ε F, B with F and B being foreground and background respectively.