Method and apparatus for multi-label segmentation
View Patent ↗A method and an apparatus for multi-label segmentation of an image are described. First an energy function is determined for the image. Then for a homogeneous region of the image variables of the energy function are grouped to a single variable. Subsequently the energy function is minimized and labels are assigned to regions of the image based on the minimized energy function.
1. A method for multi-label segmentation of an image, the method comprising:
determining an energy function for the image;
grouping variables of the energy function for a homogeneous region of the image to a single variable;
minimizing the energy function; and
assigning labels to regions of the image based on the minimized energy function.
2. The method according to claim 1 , wherein the energy function is minimized using a maximum flow algorithm.
3. The method according to claim 1 , wherein Dempster's theory of evidence is used to determine which variables and which pixels of the image are grouped.
4. The method according to claim 3 , wherein an edge weight function is defined between neighboring pixels of the image based on their appearance similarity and their task specific similarity.
5. The method according to claim 4 , wherein the appearance similarity and the task specific similarity are combined using Dempster's rule of combination.
6. The method according to claim 4 , wherein two variables of the energy function are grouped if the maximum weight of all edges connected to the two variables is smaller than a threshold.
7. The method according to claim 6 , wherein two variables of the energy function are also grouped if the value of the edge connecting the two variables or the maximum value of all edges connected to the two variables is smaller than the minimal internal difference of the two variables.
8. The method according to claim 7 , wherein the minimal internal difference of the two variables is defined as the maximum value of the minimal spanning tree plus a small value.
9. The method according to claim 1 , wherein Dempster's theory of evidence is used to determine which variables or which pixels of the image are grouped.
10. An apparatus for multi-label segmentation of an image, wherein the apparatus is:
determining an energy function for the image;
grouping variables of the energy function for a homogeneous region of the image to a single variable;
minimizing the energy function; and
assigning labels to regions of the image based on the minimized energy function.