System, computer readable storage medium, and method for segmentation and enhancement of brain mri images

Patent metadata

Field	Value
Publication number	US-2022405544-A1
Application number	US-202117336824-A
Country	US
Kind code	A1
Filing date	Jun 2, 2021
Priority date	Jun 2, 2021
Publication date	Dec 22, 2022
Grant date	—

Abstract

A system and method of 3-D image segmentation of brain images includes obtaining a 3-D MRI image, an employee phase including performing search cycles of generating solutions in a neighborhood, taking into account (a) movement of a bee's current location toward a mean value of a positive direction of a global best location and a positive direction of its own best location, (b) movement of the bee's current location toward the mean value of the positive direction of its own best location and a negative direction of the global best location, and (c) a random number, calculating a fitness value for the solutions based on membership values of pixels and distances between the pixels to cluster centers of pixels until search ends. Image segmentation of the image is performed using centers of clusters.

First claim

1 . A method of 3-D image segmentation by processing circuitry, the method comprising:
obtaining at least one 3-D Magnetic Resonance Image (MRI) image having a plurality of pixels; an employee phase including performing at least one search cycle comprising:
generating a plurality of solutions in a neighborhood of an employed bee's current location, taking into account:
(a) movement of the employed bee's current location toward a mean value of a positive direction of a global best location and a positive direction of its own best location, wherein the bee's current location is a location of a pixel of the 3-D MRI image and the global best location is a center of a cluster,
(b) movement of the employed bee's current location toward the mean value of the positive direction of its own best location and a negative direction of the global best location, and
(c) a random number;

calculating a fitness value for each of the plurality of solutions based on membership values and distances to the cluster centers, wherein each of the membership values is determined based on a degree of membership of the pixel location of the 3-D MRI image to a cluster, and the distances are distances between the pixel locations and the cluster centers; and
evaluating the solutions based on the fitness value to determine an end of the search cycle;
performing image segmentation of the 3-D MRI image based on the centers of the plurality of clusters; and
displaying the segmented 3-D image.

2 . The method of claim 1 , wherein
the 3-D MRI image is a 3-D MRI image of a human brain, and wherein the performing the image segmentation includes segmenting the image into gray matter, white matter, and cerebral spinal fluid regions.

3 . The method of claim 2 , wherein the 3-D MRI image is obtained by filtering an original MRI image into a 3-D image of intracranial tissue.

4 . The method of claim 1 , further comprising:
an onlooker phase including
every onlooker selecting a source of food with a probability that is related to the fitness value of a food source participated by employed bees;
selecting a solution based on the probability;
generating a new solution in a neighborhood of an onlooker bee's current location based on previous mean values;
evaluating the new solution;
calculating the fitness value for the new solution based on the membership values and the distances to cluster centers, wherein each of the membership values is determined based on a degree of membership of a pixel location of the 3-D MRI image to a cluster, and the distances are distances between the pixel locations and the cluster centers; and
evaluating the new solutions based on the fitness value.

5 . The method of claim 4 , further comprising:
a scout phase in which at the end of a search cycle, when a number of trials reach a limit, a food source is abandoned by an employed bee and scouts begin a new search to find new solutions randomly.

6 . The method of claim 1 , wherein in the employed phase,
the random number is [0, C], where C is a positive constant number, wherein when C increases from zero to a suitable value, a balance between exploitation and exploitation is improved, and values of C are limited to a maximum value in order to prevent a relatively weak exploration ability.

7 . A system for 3-D image segmentation, the system comprising:
processing circuitry configured to obtain at least one 3-D Magnetic Resonance Image (MRI) image having a plurality of pixels, perform, in an employee phase, at least one search cycle comprising: generating a plurality of solutions in a neighborhood of an employed bee's current location, taking into account (a) movement of the employed bee's current location toward a mean value of a positive direction of a global best location and a positive direction of its own best location, wherein the bee's current location is a location of a pixel of the 3-D MRI image and the global best location is a center of a cluster, (b) movement of the employed bee's current location toward the mean value of the positive direction of its own best location and a negative direction of the global best location, and (c) a random number, calculating a fitness value for each of the plurality of solutions based on membership values and distances to the cluster centers, wherein each of the membership values is determined based on a degree of membership of the pixel location of the 3-D MRI image to a cluster, and the distances are distances between the pixel locations and the cluster centers, and evaluating the solution based on the fitness value to determine an end of the search cycle, perform image segmentation of the 3-D MRI image based on centers of the plurality of clusters; and a display device displaying the segmented 3-D image.

8 . The system of claim 7 , wherein
the 3-D MRI image is a 3-D MRI image of a human brain, and wherein in the perform the image segmentation function the processing circuitry is further configured to segment the image into gray matter, white matter, and cerebral spinal fluid regions.

9 . The system of claim 8 , wherein the 3-D MRI image is obtained by filtering, by the processing circuitry, an original MRI image into a 3-D image of intracranial tissue.

10 . The system of claim 7 , wherein the processing circuitry is further configured to:
in an onlooker phase, every onlooker selects a source of food with a probability that is related to the fitness value of a food source participated by employed bees, selects a solution based on the probability, generates a new solution in a neighborhood of an onlooker bee's current location based on previous mean values, evaluates the new solution, calculates the fitness value for the new solution based on the membership values and the distances to cluster centers, wherein each of the membership values is determined based on a degree of membership of a pixel location of the 3-D MRI image to a cluster, and the distances are distances between the pixel locations and the cluster centers, and evaluates the new solutions based on the fitness value.

11 . The system of claim 10 , wherein the processing circuitry is further configured, in a scout phase in which at the end of each search cycle, when a number of trials reach a limit, a food source is abandoned by an employed bee and scouts begin a new search to find new solutions randomly.

12 . The system of claim 7 , wherein the processing circuitry is further configured, in the employed phase,
the random number is [0, C], where C is a positive constant number, wherein when C increases from zero to a suitable value, a balance between exploitation and exploitation is improved, and values of C are limited to a maximum value in order to prevent a relatively weak exploration ability.

13 . A non-transitory computer readable storage medium storing processing instructions, which when performed by processing circuitry, performs 3-D image segmentation comprising steps of:
obtaining at least one 3-D Magnetic Resonance Image (MRI) image having a plurality of pixels; an employee phase including performing at least one search cycle comprising: generating a plurality of solutions in a neighborhood of an employed bee's current location, taking into account (a) movement of the employed bee's current location toward a mean value of a positive direction of a global best location and a positive direction of its own best location, wherein the bee's current location is a location of a pixel of the 3-D MRI image and the global best location is a center of a cluster, (b) movement of the employed bee's current location toward the mean value of

Assignees

• Univ Imam Abdulrahman Bin Faisal

Inventors

• Alomoush Waleed Khamees Ali
• Alrosan Ayat Ahmad Meqbel

Classifications

• G06T2207/30016

  Brain · CPC title

• G06T7/143

  involving probabilistic approaches, e.g. Markov random field [MRF] modelling · CPC title

• G06T7/11 Primary

  Region-based segmentation · CPC title

• G06T2207/10088

  Magnetic resonance imaging [MRI] · CPC title

• G06T7/174

  involving the use of two or more images · CPC title