Automatic 3D segmentation and cortical surfaces reconstruction from T1 MRI

Patent metadata

Field	Value
Publication number	US-9530206-B2
Application number	US-201514675265-A
Country	US
Kind code	B2
Filing date	Mar 31, 2015
Priority date	Mar 31, 2015
Publication date	Dec 27, 2016
Grant date	Dec 27, 2016

Abstract

An apparatus and method for performing automatic 3D image segmentation and reconstruction of organ structures, which is particularly well-suited for use on cortical surfaces is presented. A brain extraction process removes non-brain image elements, then classifies brain tissue as to type in preparation for a cerebrum segmentation process that determines which portions of the image information belong to specific physiological structures. Ventricle filling is performed on the image data based on information from a ventricle extraction process. A reconstruction process follows in which specific surfaces, such as white matter (WM) and grey matter (GM), are reconstructed.

First claim

What is claimed is:

1. An apparatus for automatically processing medical images of brain regions, comprising:
(a) a medical imaging system configured for scanning a patient's brain in three dimensions to generate three dimensional (3D) medical brain image data;
(b) at least one processor configured for automatically performing image processing on said 3D medical brain image data;
(c) a memory storing instructions executable by said processor; and
(d) a display device coupled to said computer processor, said display device configured for displaying 3D medical images;
(e) said instructions when executed performing steps comprising:
(i) determining brain structures in a brain extraction process which removes non-brain image elements;
(ii) classifying brain tissue types;
(iii) segmenting cerebrum imaging which determines which portions of the image information belong to specific physiological structures;
(iv) anatomical refinement comprising ventricle filling in response to ventricle information extracted from the image, removal of dura mater, and correcting subcortical area;
(v) reconstructing surfaces of the brain including both white and grey matter in response to generating a surface mesh and correcting mesh topology;
(vi) wherein each of the above steps are performed in response to said programming without user interaction or intervention; and
(vii) displaying a 3D brain image showing classification of brain tissues, segmentation of different brain portions and surface reconstructions, with ventricle filling on said display.

2. The apparatus recited in claim 1 , wherein said three dimensional (3D) medical image data comprises a form of magnetic resonance images (MRI).

3. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for performing said segmentation and reconstruction of cortical surfaces which includes reconstructing both inner-surfaces and outer-surfaces of the cortex.

4. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for brain extraction processing based on stored anatomical information.

5. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for brain extraction processing to remove non-brain voxels in the head region, including comprising skin, bone, muscle, fat and dura.

6. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for performing said brain extraction processing including non-uniformity correction which corrects non-uniformities in raw magnetic imaging data due to magnetic field inhomogeneity.

7. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for separating cerebellum and brain stem image information from cerebrum specific physiological structures in the process of performing said segmenting cerebrum imaging.

8. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for said segmenting cerebrum imaging, comprising:
(a) identifying mid-sagittal plane (MSP) in response to a process of searching for a plane in medical image space about which maximal symmetry is found;
(b) identifying corpus callosum and pons from intersection of MSP and white matter (WM) member function as output under tissue classification; and
(c) utilizing the identity of one set of structures for identifying other sets of structures.

9. The apparatus recited in claim 8 , wherein said instructions for said processor are configured for performing said segmenting cerebrum imaging in which the identity of corpus callosum and pons are utilized as initial seeds of cerebrum and cerebellum and brain stem identification with a graph-cut being applied for segmenting cerebrum imaging.

10. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for said ventricle filling and performed in response to ventricle information received from multiple orientations.

11. The apparatus recited in claim 10 , wherein said instructions for said processor are configured for providing multiple orientations of ventricle information as from at least a front, side and back view.

12. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for performing said ventricle filling in which internal cavities of the white matter (WM) are filled before performing the surface reconstruction toward preserving the physical shape of the cortex.

13. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for preserving the physical shape of the cortex, estimated as a spherical topology in response to use of Euler characteristics.

14. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for performing said surface reconstruction of the white matter (WM) by generating an isosurface mesh from WM volume data.

15. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for performing said surface reconstruction of grey matter (GM) based on WM segmentation results to refine GM and cerebrospinal fluid (CSF) classification to improve sulci description, followed by generating a surface mesh.

16. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for performing said reconstruction which includes a sulci refinement process during GM surface reconstruction.

17. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for reconstruction of brain surfaces by outputting image data which is in different colors denoting different regions of the brain.

18. The apparatus recited in claim 1 , wherein said instructions for said processor are configured for reconstruction of brain surfaces while preserving physical shape of the cortex, as it is a single sheet of connected tissue, whereby said programming is configured for closing this surface at the brain stem.

19. An apparatus for automatically processing medical images of brain regions, comprising:
(a) a medical imaging system configured for scanning a patient's brain in three dimensions to generate three dimensional (3D) medical brain image data;
(b) a computer processor configured for automatically performing image processing on said 3D medical brain image data;
(c) a display device coupled to said computer processor, said display device configured for displaying 3D medical images; and
(d) a memory storing programming executable on the computer processor;
(e) said programming when executed performing steps comprising:
(i) determining brain structures in a brain extraction process which removes non-brain image elements;
(ii) classifying brain tissue types;
(iii) segmenting cerebrum imaging which determines which portions of the image information belong to specific physiological structures, in response to: (iii)(A) identifying mid-sagittal plane (MSP) in response to a process of searching for a plane in medical image space about which maximal symmetry is found, (iii)(B) identifying corpus callosum and pons from intersection of MSP and white matter (WM) member function as output under tissue classification, and (iii)(C) the identifying of one set of structures then utilized in identifying others sets of structures;
(iv) anatomical refinement comprising ventricle filling in response to ventricle information extracted from the image, removal of dura mater, and correcting subcortical area;
(v) reconstruction o

Assignees

• Sony Corp

Inventors

• Liu Ming-Chang
• Song Bi

Classifications

• G06T7/11 Primary

  Region-based segmentation · CPC title

• G06T12/00 Primary

  Tomographic reconstruction from projections · CPC title

• G06F18/24

  Classification techniques · CPC title

• G06T7/68

  of symmetry · CPC title

• G06T7/0012 Primary

  Biomedical image inspection · CPC title