# US 2018/0025488 A1 · Automatic 3d brain tumor segmentation and classification

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2018/0025488 A1
- Document ID: US20180025488A1
- Status: Published application
- Kind: A1
- Application number: 15/218,986
- Filed: 2016-07-25
- Published: 2018-01-25
- Earliest claimed: 2016-07-25
- Assignees: Sony Corporation
- Applicants: Sony Corporation
- Inventors: Chen-Rui Chou; Bi Song; Ming-Chang Liu
- Web page: https://patentgrep.com/patent/US20180025488A1

## CPC

- G06T 7/0012: {Biomedical image inspection}
- A61B 5/055: involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging
- A61B 5/7264: {Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems}
- G01R 33/4808: {Multimodal MR, e.g. MR combined with positron emission tomography [PET], MR combined with ultrasound or MR combined with computed tomography [CT]}
- G01R 33/4828: {Resolving the MR signals of different chemical species, e.g. water-fat imaging}
- G06F 18/24: Classification techniques
- G06F 18/24323: {Tree-organised classifiers}
- G06T 7/10: Segmentation; Edge detection (motion-based segmentation G06T7/215)
- G06T 7/11: Region-based segmentation
- G06T 7/12: Edge-based segmentation
- G06T 7/187: involving region growing; involving region merging; involving connected component labelling
- G06V 10/764: using classification, e.g. of video objects
- A61B 2576/00: Medical imaging apparatus involving image processing or analysis (A61B1/00009, A61B6/52 and A61B8/52 take precedence)
- G06T 2200/04: involving 3D image data
- G06T 2207/10088: Magnetic resonance imaging [MRI]
- G06T 2207/20081: Training; Learning
- G06T 2207/20156: Automatic seed setting
- G06T 2207/30016: Brain
- G06T 2207/30096: Tumor; Lesion
- G06V 2201/03: Recognition of patterns in medical or anatomical images
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- G01R: MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES (indicating correct tuning of resonant circuits H03J3/12)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

A fully automatic brain tumor segmentation and classification method and system improve the healthcare experience with machine intelligence. The automatic brain tumor segmentation and classification method and system utilize whole tumor segmentation and multi-class tumor segmentation to provide accurate analysis.

## Claims

### Claim 1

A method programmed in a non-transitory memory of a device comprising:
performing whole tumor segmentation, wherein the whole tumor segmentation includes:
utilizing multi-modal MRIs including T1, T1c, Flair and T2;
determining a least-tumor hemisphere;
classifying tissue;
identifying intensity control points based on intensities of different tissue structures;
normalizing the intensity for each of the MRIs based on the intensity control points;
locating initial tumor seeds; and
segmenting a tumor based on the tumor seeds; and
performing multi-class tumor segmentation.

### Claim 2 (depends on 1)

The method of claim 1 wherein the whole tumor segmentation includes: data normalization and initial segmentation.

### Claim 3

(canceled)

### Claim 4 (depends on 1)

The method of claim 1 wherein the multi-class tumor segmentation includes: feature extraction, voxel classification and refinement.

### Claim 5 (depends on 4)

The method of claim 4 wherein feature extraction includes: determining voxel-wise features and context features, wherein voxel-wise features include appearance features, texture features and location features.

### Claim 6 (depends on 5)

The method of claim 5 wherein context feature extraction further includes: extracting a mean intensity in each octant of an MRI, extracting multiscale context features and combining the context features from a T1c MRI and a T2 MRI.

### Claim 7 (depends on 6)

The method of claim 6 wherein voxel classification utilizes information from the feature extraction and decision trees to classify a tumor.

### Claim 8 (depends on 7)

The method of claim 7 wherein refinement includes pathology-guided refinement to ensure a correct classification of the tumor.

### Claim 9

An apparatus comprising:
a non-transitory memory for storing an application, the application for:
performing whole tumor segmentation, wherein the whole tumor segmentation includes:
utilizing multi-modal MRIs including T1, T1c, Flair and T2;
determining a least-tumor hemisphere;
classifying tissue;
identifying intensity control points based on intensities of different tissue structures;
normalizing the intensity for each of the MRIs based on the intensity control points;
locating initial tumor seeds; and
segmenting a tumor based on the tumor seeds; and
performing multi-class tumor segmentation; and
a processing component coupled to the memory, the processing component configured for processing the application.

### Claim 10 (depends on 9)

The apparatus of claim 9 wherein the whole tumor segmentation includes: data normalization and initial segmentation.

### Claim 11

(canceled)

### Claim 12 (depends on 9)

The apparatus of claim 9 wherein the multi-class tumor segmentation includes: feature extraction, voxel classification and refinement.

### Claim 13 (depends on 12)

The apparatus of claim 12 wherein feature extraction includes: voxel-wise features and context features, wherein voxel-wise features include determining appearance features, texture features and location features.

### Claim 14 (depends on 13)

The apparatus of claim 13 wherein context feature extraction further includes: extracting a mean intensity in each octant of an MRI, extracting multiscale context features and combining the context features from a T1c MRI and a T2 MRI.

### Claim 15 (depends on 14)

The apparatus of claim 14 wherein voxel classification utilizes information from the feature extraction and decision trees to classify a tumor.

### Claim 16 (depends on 15)

The apparatus of claim 15 wherein refinement includes pathology-guided refinement to ensure a correct classification of the tumor.

### Claim 17

A system comprising:
a magnetic resonance imaging device; and
a computing device configured for:
performing whole tumor segmentation, wherein the whole tumor segmentation includes:
utilizing multi-modal MRIs including T1, T1c, Flair and T2;
determining a least-tumor hemisphere;
classifying tissue;
identifying intensity control points based on intensities of different tissue structures:
normalizing the intensity for each of the MRIs based on the intensity control points;
locating initial tumor seeds; and
segmenting a tumor based on the tumor seeds; and
performing multi-class tumor segmentation.

### Claim 18 (depends on 17)

The system of claim 17 wherein the whole tumor segmentation includes: data normalization and initial segmentation.

### Claim 19

(canceled)

### Claim 20 (depends on 17)

The system of claim 17 wherein the multi-class tumor segmentation includes: feature extraction, voxel classification and refinement.

### Claim 21 (depends on 20)

The system of claim 20 wherein feature extraction includes: voxel-wise features and context features, wherein voxel-wise features include determining appearance features, texture features and location features.

### Claim 22 (depends on 21)

The system of claim 21 wherein context feature extraction further includes: extracting a mean intensity in each octant of an MRI, extracting multiscale context features and combining the context features from a T1c MRI and a T2 MRI.

### Claim 23 (depends on 22)

The system of claim 22 wherein voxel classification utilizes information from the feature extraction and decision trees to classify a tumor.

### Claim 24 (depends on 23)

The system of claim 23 wherein refinement includes pathology-guided refinement to ensure a correct classification of the tumor.

### Claim 25

A method programmed in a non-transitory memory of a device comprising:
performing whole tumor segmentation; and
performing multi-class tumor segmentation, wherein the multi-class tumor segmentation includes: feature extraction, voxel classification and refinement, wherein feature extraction includes: determining voxel-wise features and context features, wherein voxel-wise features include appearance features, texture features and location features, wherein context feature extraction further includes: extracting a mean intensity in each octant of an MRI, extracting multiscale context features and combining the context features from a T1c MRI and a T2 MRI.

### Claim 26 (depends on 25)

The method of claim 25 wherein the whole tumor segmentation includes: data normalization and initial segmentation.

### Claim 27 (depends on 25)

The method of claim 25 wherein voxel classification utilizes information from the feature extraction and decision trees to classify a tumor.

### Claim 28 (depends on 27)

The method of claim 27 wherein refinement includes pathology-guided refinement to ensure a correct classification of the tumor.

### Claim 29

An apparatus comprising:
a non-transitory memory for storing an application, the application for:
performing whole tumor segmentation; and
performing multi-class tumor segmentation, wherein the multi-class tumor segmentation includes: feature extraction, voxel classification and refinement, wherein feature extraction includes: determining voxel-wise features and context features, wherein voxel-wise features include appearance features, texture features and location features, wherein context feature extraction further includes: extracting a mean intensity in each octant of an MRI, extracting multiscale context features and combining the context features from a T1c MRI and a T2 MRI; and
a processing component coupled to the memory, the processing component configured for processing the application.

### Claim 30 (depends on 29)

The apparatus of claim 29 wherein the whole tumor segmentation includes: data normalization and initial segmentation.

### Claim 31 (depends on 29)

The apparatus of claim 29 wherein voxel classification utilizes information from the feature extraction and decision trees to classify a tumor.

### Claim 32 (depends on 31)

The apparatus of claim 31 wherein refinement includes pathology-guided refinement to ensure a correct classification of the tumor.