CN114092425A - Cerebral ischemia scoring method based on diffusion weighted image, electronic device and medium - Google Patents
Cerebral ischemia scoring method based on diffusion weighted image, electronic device and medium Download PDFInfo
- Publication number
- CN114092425A CN114092425A CN202111336632.1A CN202111336632A CN114092425A CN 114092425 A CN114092425 A CN 114092425A CN 202111336632 A CN202111336632 A CN 202111336632A CN 114092425 A CN114092425 A CN 114092425A
- Authority
- CN
- China
- Prior art keywords
- image
- blood supply
- brain
- supply area
- images
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000009792 diffusion process Methods 0.000 title claims abstract description 47
- 201000006474 Brain Ischemia Diseases 0.000 title claims abstract description 23
- 206010008120 Cerebral ischaemia Diseases 0.000 title claims abstract description 21
- 206010008118 cerebral infarction Diseases 0.000 title claims abstract description 21
- 238000013077 scoring method Methods 0.000 title claims description 14
- 230000036770 blood supply Effects 0.000 claims abstract description 81
- 210000004556 brain Anatomy 0.000 claims abstract description 77
- 238000013145 classification model Methods 0.000 claims abstract description 31
- 230000000302 ischemic effect Effects 0.000 claims abstract description 27
- 238000000034 method Methods 0.000 claims abstract description 24
- 238000012545 processing Methods 0.000 claims abstract description 12
- 238000012549 training Methods 0.000 claims description 16
- 238000004590 computer program Methods 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 7
- 230000002452 interceptive effect Effects 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 230000009467 reduction Effects 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 4
- 238000004422 calculation algorithm Methods 0.000 claims description 3
- 230000004044 response Effects 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 230000005855 radiation Effects 0.000 abstract description 6
- 238000005516 engineering process Methods 0.000 abstract description 5
- 238000013135 deep learning Methods 0.000 abstract description 3
- 208000032382 Ischaemic stroke Diseases 0.000 description 11
- 208000006011 Stroke Diseases 0.000 description 9
- 238000002597 diffusion-weighted imaging Methods 0.000 description 9
- 230000008569 process Effects 0.000 description 9
- 230000033001 locomotion Effects 0.000 description 6
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 6
- 238000003384 imaging method Methods 0.000 description 5
- 210000005013 brain tissue Anatomy 0.000 description 4
- 208000028867 ischemia Diseases 0.000 description 4
- 206010008190 Cerebrovascular accident Diseases 0.000 description 3
- 238000005481 NMR spectroscopy Methods 0.000 description 3
- 230000002490 cerebral effect Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- 238000007792 addition Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 201000010099 disease Diseases 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- 239000006185 dispersion Substances 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- 230000001575 pathological effect Effects 0.000 description 2
- 238000004393 prognosis Methods 0.000 description 2
- 210000003625 skull Anatomy 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 206010061216 Infarction Diseases 0.000 description 1
- 102000007999 Nuclear Proteins Human genes 0.000 description 1
- 108010089610 Nuclear Proteins Proteins 0.000 description 1
- 230000001154 acute effect Effects 0.000 description 1
- 210000001367 artery Anatomy 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 210000001159 caudate nucleus Anatomy 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000007574 infarction Effects 0.000 description 1
- 238000007917 intracranial administration Methods 0.000 description 1
- 230000002427 irreversible effect Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 210000003141 lower extremity Anatomy 0.000 description 1
- 210000002637 putamen Anatomy 0.000 description 1
- 230000000542 thalamic effect Effects 0.000 description 1
- 230000002537 thrombolytic effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20032—Median filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30016—Brain
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
According to the method, each blood supply area of the brain is automatically divided into images based on the brain diffusion weighted images of the patient, whether each blood supply area image of the patient has an ischemic focus or not is respectively classified by using each blood supply area image classification model based on VGG-16 to obtain the classification result of each blood supply area image, and the cerebral ischemia score of the patient is obtained according to the number of the ischemic focuses or the number of the ischemic focuses in each blood supply area image according to the classification result. According to the method, the full-automatic grading of the cerebral ischemia degree is realized by means of a deep learning image processing technology, the grading based on the brain diffusion weighted image does not cause radiation damage to a patient, the experience requirements on doctors are reduced, and the efficiency is improved.
Description
Technical Field
The invention relates to a medical image processing technology, in particular to a cerebral ischemia scoring method based on diffusion weighted images, electronic equipment and a medium.
Background
Cerebral apoplexy is the first cause of death in China at present, and Acute Ischemic Stroke (AIS) is the most common type of Stroke and accounts for about 70-80% of the total incidence rate of cerebral apoplexy. Acute ischemic stroke is a disease with high disability rate and high mortality rate, and more than 40% of AIS is caused by intracranial acute large vessel occlusion. If AIS occurs, irreversible damage to the brain tissue of the patient can occur, seriously compromising the life and health of the patient. The key of AIS treatment is the time window, carries out timely, effectual treatment to the cerebral apoplexy patient at the early stage of disease, can effectively reduce final infarct area, saves more brain tissues to reduce the disability rate. Therefore, the skull image of the patient is timely acquired at the early stage of AIS, and accurate diagnosis and evaluation are carried out aiming at the skull image, so that the important value is provided for the early treatment and prognosis of the patient.
The Early stage CT Score (ASPECTS) of Alberta Stroke Program Early CT Score is a simple and reliable method for evaluating the Early ischemic change of an artery blood supply area in the brain of a patient suffering from ischemic Stroke, and is helpful for judging the thrombolytic effect and the long-term prognosis. The ASPECTS score has proven to be a powerful predictor for diagnosing the status of stroke patients. However, the prior art ASPECTS scoring method has at least the following two problems: (1) the ASPECTS score is evaluated based on CT images, which cause X-ray radiation damage to the patient; (2) the ASPECTS score for patients with ischemic stroke is mainly dependent on the experience of doctors, and the doctors with high experience need to observe the ischemia condition of each blood supply area in the CT images of the brain of the patients to score.
Diffusion-Weighted Imaging (DWI) is an Imaging method which reflects the irregular thermal motion condition of water molecules in a living body noninvasively at a molecular level, is a scanning Imaging method through nuclear magnetic resonance equipment, does not cause radiation damage to a patient in the process, and is a common method for brain examination. The stroke limits the free movement of water molecules in normal brain tissues, and DWI is very sensitive to the dispersion movement change of the water molecules, so that the pathological and physiological changes in the stroke process can be reflected. When ischemic changes occur early, DWI can more accurately detect subtle signs of ischemia in a shorter time, and can be used for assessing ischemic penumbra in AIS patients over a time window.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a cerebral ischemia scoring method based on diffusion weighted images, electronic equipment and a medium.
In a first aspect, a brain ischemia scoring method based on diffusion-weighted images is provided, which includes:
acquiring a brain diffusion weighted image of a patient, wherein the brain diffusion weighted image is obtained by scanning the brain of the patient by using magnetic resonance equipment;
preprocessing the brain diffusion weighted image to obtain a spatially normalized brain diffusion weighted image;
acquiring mask images respectively corresponding to 10 blood supply areas of a brain, wherein each mask image respectively comprises position information of each blood supply area;
obtaining images of each blood supply area of the patient according to each mask image and the brain diffusion weighted image of the spatial standardization;
classifying whether each blood supply area image of the patient has an ischemic focus or not by using each blood supply area image classification model based on VGG-16 to respectively obtain a classification result of each blood supply area image;
and obtaining the cerebral ischemia score of the patient according to the number of ischemic foci or ischemic foci obtained by the classification result in each blood supply area image.
In one embodiment, the pre-processing the brain diffusion weighted image comprises:
carrying out noise reduction processing on the brain diffusion weighted image by utilizing a nonlinear smooth median filtering algorithm;
and registering the brain diffusion weighted image subjected to noise reduction processing with a standard brain template image on the MNI152 space to obtain the brain diffusion weighted image subjected to spatial standardization.
In one embodiment, the obtaining images of the respective donor areas of the patient from the respective mask images and the spatially normalized brain diffusion weighted image comprises:
and respectively carrying out matrix multiplication on the mask images and the brain diffusion weighted images with the space standardization to obtain images of each blood supply area of the patient.
In one embodiment, the mask images corresponding to the 10 blood supply regions of the brain are obtained by:
in response to an interactive instruction input by a professional radiologist through an interactive interface, drawing position ranges corresponding to 10 blood supply areas of the brain respectively on a standard brain template image in the MNI152 space;
and setting the pixel point of the specific blood supply area on the standard brain template image as 1 and setting other pixel points as 0 for each specific blood supply area, and respectively generating a mask image corresponding to each specific blood supply area.
In one embodiment, the VGG-16-based image classification model of each blood supply region is obtained by the following steps:
respectively constructing an initial image classification model based on VGG-16 for each blood supply area;
respectively collecting a plurality of training images with focuses and a plurality of training images without focuses as training data to train the initial image classification model aiming at each blood supply area, and obtaining the trained image classification model;
and verifying each trained image classification model by using verification data which is not coincident with the training data to obtain a final blood supply area image classification model.
In a second aspect, the present invention provides an electronic device comprising:
a memory storing a computer program;
and the processor is in communication connection with the memory and executes the cerebral ischemia scoring method based on diffusion weighted images when the computer program is called.
In a third aspect, the present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed, implements a diffusion-weighted image-based cerebral ischemia scoring method provided by the present invention.
Compared with the prior art, the brain diffusion weighting method has the advantages that each blood supply area of the brain is automatically divided into the images based on the brain diffusion weighting images of the patient, each blood supply area image classification model based on VGG-16 is used for classifying whether each blood supply area image of the patient has an ischemic focus or not to obtain the classification result of each blood supply area image, and the cerebral ischemia score of the patient is obtained according to the number of the ischemic focuses or the number of the ischemic focuses of the classification result in each blood supply area image. According to the method, the full-automatic grading of the cerebral ischemia degree is realized by means of a deep learning image processing technology, the grading based on the brain diffusion weighted image does not cause radiation damage to a patient, the experience requirements on doctors are reduced, and the efficiency is improved.
Drawings
FIG. 1 is a flow chart of a cerebral ischemia scoring method based on diffusion-weighted images according to the present invention;
FIG. 2 is a schematic diagram of 10 blood supply regions in the brain of a human;
FIG. 3 is an embodiment of a VGG-16-based blood supply region image classification model structure;
FIG. 4 is an embodiment of a loss variation curve of a blood supply region image classification model corresponding to the M2 blood supply region in the training process.
Detailed Description
The objects, technical solutions and advantages of the present invention have been described in further detail with reference to the preferred embodiments, it should be understood that the above description is only illustrative of the preferred embodiments of the present invention, and should not be construed as limiting the present invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention should be included in the scope of the present invention.
Example 1
As shown in fig. 1, a brain ischemia scoring method based on diffusion-weighted images is provided, which includes:
and step S10, acquiring a brain diffusion weighted image of the patient, wherein the brain diffusion weighted image is obtained by scanning the brain of the patient by using a magnetic resonance device.
Diffusion-Weighted Imaging (DWI) is an Imaging method which reflects the irregular thermal motion condition of water molecules in a living body noninvasively at a molecular level, is a scanning Imaging method through nuclear magnetic resonance equipment, does not cause radiation damage to a patient in the process, and is a common method for brain examination. The stroke limits the free movement of water molecules in normal brain tissues, and DWI is very sensitive to the dispersion movement change of the water molecules, so that the pathological and physiological changes in the stroke process can be reflected. When ischemic changes occur early, DWI can more accurately detect subtle signs of ischemia in a shorter time, and can be used for assessing ischemic penumbra in AIS patients over a time window.
And step S20, preprocessing the brain diffusion weighted image to obtain a spatially normalized brain diffusion weighted image.
In one embodiment, the pre-processing the brain diffusion weighted image comprises:
s201, noise reduction processing is carried out on the brain diffusion weighted image by utilizing a nonlinear smooth median filtering algorithm. In order to remove the pulse interference of the magnetic resonance image in the acquisition process, the DWI image of the patient is preprocessed by nonlinear smooth median filtering, so that the influence of various physiological noises is avoided
S202, registering the brain diffusion weighted image subjected to noise reduction processing with a standard brain template image on the MNI152 space to obtain the brain diffusion weighted image subjected to spatial standardization. Because different nuclear magnetic resonance equipment has different scanning imaging parameters, different positions of patients lying on the magnetic resonance equipment and certain differences of brain forms of different patients, in order to realize the automatic division of the brain blood supply areas by using a computer image processing technology, the brain DWI images of different patients are registered on the standard brain template image on the MNI152 space, the brain DWI image space standardization of different patients is realized, and the division of the brain blood supply areas is realized by using mask images of 10 blood supply areas on the MNI152 space in the follow-up process.
Step S30, mask images corresponding to the 10 blood supply regions of the brain are obtained, and each mask image includes position information of each blood supply region.
Fig. 2 is a schematic diagram of 10 blood supply regions of a human brain, which includes: 7 regions of the nucleolar layer (i.e., thalamic and striatal planes) M1, M2, M3, islet leaves I, putamen L, caudate nucleus C, and inner capsular hindlimb IC; and 3 regions at the above-mentioned level of the nuclei (2 cm at the level of the nuclei), M4, M5 and M6, respectively.
In one embodiment, mask images corresponding to 10 blood supply regions of the brain are obtained by:
s301, in response to an interactive instruction input by a professional radiologist through an interactive interface, drawing position ranges corresponding to 10 brain blood supply areas on a standard brain template image in MNI152 space;
s302, setting the pixel point of the specific blood supply area on the standard brain template image as 1 and setting other pixel points as 0 aiming at each specific blood supply area, and respectively generating a mask image corresponding to each specific blood supply area. It will be appreciated that any digital image that can be processed by a computer, whether a two-dimensional planar image or a three-dimensional volumetric image, can be equated with a mathematical matrix, and that the processing of the image can be achieved by means of matrix calculations. MASK image (MASK) refers to an image matrix consisting of 0 and 1, whose roles are: and performing matrix multiplication operation on the MASK image with the same size and the image to be processed, wherein the position of the image to be processed, which corresponds to the MASK with 0, is shielded, and the position with 1 is reserved after the image to be processed is processed.
Step S40, obtaining images of each blood supply area of the patient from each mask image and the spatially normalized brain diffusion weighted image.
In one embodiment, the step S40 is specifically implemented as follows: and respectively carrying out matrix multiplication on the mask images and the brain diffusion weighted images with the space standardization to obtain images of each blood supply area of the patient.
And step S50, classifying whether each blood supply area image of the patient has an ischemic focus or not by using each blood supply area image classification model based on VGG-16, and obtaining the classification result of each blood supply area image.
In one embodiment, the VGG-16-based image classification model of each blood supply region is obtained by the following steps:
s501, aiming at each blood supply area, an initial image classification model based on VGG-16 is respectively constructed. FIG. 3 shows an embodiment of an initial image classification model structure based on VGG-16. In one embodiment, the specific neural network structure may be: and taking the pooling layer as a boundary, the VGG-16 has 6 block structures, and the number of channels in each block structure is the same. VGG-16 has 16 layers in total, wherein the convolutional layer is 13 layers, the full connection is 3 layers, and the pooling layer does not relate to weight. For the VGG-16 convolutional neural network, 13 convolutional layers and 5 pooling layers are responsible for extracting features, and the last 3 fully-connected layers are responsible for completing classification tasks.
S502, aiming at each blood supply area, respectively collecting a plurality of training images with focuses and a plurality of training images without focuses as training data to train the initial image classification model, and obtaining the trained image classification model.
S503, verifying each trained image classification model by using verification data which is not overlapped with the training data to obtain a final blood supply area image classification model.
The trained blood supply area image classification model can realize the classification of the blood supply area image obtained in the step S40. In one embodiment, the classification result is output by each blood supply region image classification model in the following manner: the output value is 1 if there is no ischemic focus in the blood supply region image, and is 0 if there is an ischemic focus.
In one embodiment, the number of training and validation sets and the validation results for the VGG-16 based image classification model are shown in tables 1 and 2. Fig. 4 shows a Loss variation curve of the blood supply region image classification model corresponding to the M2 blood supply region in the training process, and it can be seen that the Loss decreases with the progress of the training process, which indicates that the corresponding classification accuracy is improved.
Table 1:
table 2:
and step S60, obtaining the cerebral ischemia score of the patient according to the number of the ischemic foci or the ischemic foci as the classification result in each blood supply area image. In one example, 10 ischemic foci are scored as 10 in 10 areas of the brain, with each ischemic focus area being subtracted by one, a lower score indicating more severe ischemia.
Example 2
Provided is an electronic device including:
a memory storing a computer program;
and the processor is in communication connection with the memory and executes the cerebral ischemia scoring method based on diffusion weighted images when the computer program is called.
Example 3
There is provided a computer readable storage medium having stored thereon a computer program which, when executed, implements a diffusion-weighted image-based cerebral ischemia scoring method provided by the present invention.
Compared with the prior art, the brain diffusion weighting method has the advantages that each blood supply area of the brain is automatically divided into the images based on the brain diffusion weighting images of the patient, each blood supply area image classification model based on VGG-16 is used for classifying whether each blood supply area image of the patient has an ischemic focus or not to obtain the classification result of each blood supply area image, and the cerebral ischemia score of the patient is obtained according to the number of the ischemic focuses or the number of the ischemic focuses of the classification result in each blood supply area image. According to the method, the full-automatic grading of the cerebral ischemia degree is realized by means of a deep learning image processing technology, the grading based on the brain diffusion weighted image does not cause radiation damage to a patient, the experience requirements on doctors are reduced, and the efficiency is improved.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and additions can be made without departing from the method of the present invention, and these modifications and additions should also be regarded as the protection scope of the present invention.
Claims (7)
1. A cerebral ischemia scoring method based on diffusion weighted images is characterized by comprising the following steps:
acquiring a brain diffusion weighted image of a patient, wherein the brain diffusion weighted image is obtained by scanning the brain of the patient by using magnetic resonance equipment;
preprocessing the brain diffusion weighted image to obtain a spatially normalized brain diffusion weighted image;
acquiring mask images respectively corresponding to 10 blood supply areas of a brain, wherein each mask image respectively comprises position information of each blood supply area;
obtaining images of each blood supply area of the patient according to each mask image and the brain diffusion weighted image of the spatial standardization;
classifying whether each blood supply area image of the patient has an ischemic focus or not by using each blood supply area image classification model based on VGG-16 to respectively obtain a classification result of each blood supply area image;
and obtaining the cerebral ischemia score of the patient according to the number of ischemic foci or ischemic foci obtained by the classification result in each blood supply area image.
2. The method of claim 1, wherein the pre-processing the brain diffusion weighted image comprises:
carrying out noise reduction processing on the brain diffusion weighted image by utilizing a nonlinear smooth median filtering algorithm;
and registering the brain diffusion weighted image subjected to noise reduction processing with a standard brain template image on the MNI152 space to obtain the brain diffusion weighted image subjected to spatial standardization.
3. The method of claim 2, wherein obtaining images of the patient's donor areas from the mask images and the spatially normalized brain diffusion weighted images comprises:
and respectively carrying out matrix multiplication on the mask images and the brain diffusion weighted images with the space standardization to obtain images of each blood supply area of the patient.
4. The method according to claim 3, wherein the mask images corresponding to the 10 blood supply regions of the brain are obtained by:
in response to an interactive instruction input by a professional radiologist through an interactive interface, drawing position ranges corresponding to 10 blood supply areas of the brain respectively on a standard brain template image in the MNI152 space;
and setting the pixel point of the specific blood supply area on the standard brain template image as 1 and setting other pixel points as 0 for each specific blood supply area, and respectively generating a mask image corresponding to each specific blood supply area.
5. The method of claim 1, wherein the VGG-16 based image classification models for each donor area are obtained by:
respectively constructing an initial image classification model based on VGG-16 for each blood supply area;
respectively collecting a plurality of training images with focuses and a plurality of training images without focuses as training data to train the initial image classification model aiming at each blood supply area, and obtaining the trained image classification model;
and verifying each trained image classification model by using verification data which is not coincident with the training data to obtain a final blood supply area image classification model.
6. An electronic device, characterized in that the electronic device comprises:
a memory storing a computer program;
a processor, communicatively connected to the memory, for executing the method for scoring cerebral ischemia based on diffusion-weighted images according to any one of claims 1 to 5 when the computer program is invoked.
7. A computer-readable storage medium having stored thereon a computer program, wherein the computer program when executed implements a diffusion-weighted image-based cerebral ischemia scoring method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111336632.1A CN114092425A (en) | 2021-11-12 | 2021-11-12 | Cerebral ischemia scoring method based on diffusion weighted image, electronic device and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111336632.1A CN114092425A (en) | 2021-11-12 | 2021-11-12 | Cerebral ischemia scoring method based on diffusion weighted image, electronic device and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114092425A true CN114092425A (en) | 2022-02-25 |
Family
ID=80300070
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111336632.1A Pending CN114092425A (en) | 2021-11-12 | 2021-11-12 | Cerebral ischemia scoring method based on diffusion weighted image, electronic device and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114092425A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114628036A (en) * | 2022-05-17 | 2022-06-14 | 中南大学湘雅医院 | Brain ischemia risk prediction platform based on neural network |
-
2021
- 2021-11-12 CN CN202111336632.1A patent/CN114092425A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114628036A (en) * | 2022-05-17 | 2022-06-14 | 中南大学湘雅医院 | Brain ischemia risk prediction platform based on neural network |
| CN114628036B (en) * | 2022-05-17 | 2022-08-02 | 中南大学湘雅医院 | Brain ischemia risk prediction platform based on neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7383010B2 (en) | Medical image recognition method and system, model training method, computer device, and program | |
| CN110232383B (en) | Focus image recognition method and focus image recognition system based on deep learning model | |
| CN110060774B (en) | Thyroid nodule identification method based on generative confrontation network | |
| KR101857624B1 (en) | Medical diagnosis method applied clinical information and apparatus using the same | |
| CN110853111B (en) | Medical image processing system, model training method and training device | |
| Hassan et al. | Joint segmentation and quantification of chorioretinal biomarkers in optical coherence tomography scans: A deep learning approach | |
| CN111008984A (en) | Method and system for automatically drawing contour line of normal organ in medical image | |
| CN109255354B (en) | Medical CT-oriented computer image processing method and device | |
| CN112348785B (en) | Epileptic focus positioning method and system | |
| CN112950737B (en) | Fundus fluorescence contrast image generation method based on deep learning | |
| CN110246109A (en) | Merge analysis system, method, apparatus and the medium of CT images and customized information | |
| US11883099B2 (en) | Noninvasive techniques for identifying choroidal neovascularization in retinal scans | |
| US20220012883A1 (en) | Learning method for generating multiphase collateral image and multiphase collateral image generating method using maching learning | |
| WO2019169403A1 (en) | Neural network classification of osteolysis and synovitis near metal implants | |
| Dong et al. | Identifying carotid plaque composition in MRI with convolutional neural networks | |
| KR20190087681A (en) | A method for determining whether a subject has an onset of cervical cancer | |
| CN114092425A (en) | Cerebral ischemia scoring method based on diffusion weighted image, electronic device and medium | |
| CN116030063B (en) | Classification diagnosis system, method, electronic device and medium for MRI image | |
| CN111754452A (en) | Detection method, medium and terminal for lower limb deep venous thrombosis based on deep learning | |
| CN115578285B (en) | Mammary gland molybdenum target image detail enhancement method and system | |
| CN112884759B (en) | Method and related device for detecting metastasis state of axillary lymph nodes of breast cancer | |
| CN112766333B (en) | Medical image processing model training method, medical image processing method and device | |
| Ahamad et al. | Brain tumor detection using deep learning | |
| Junaid et al. | Data Science is all set to revolutionize the Ultrasound diagnosis in medical health care | |
| deManizales et al. | to deep learninguniversidad appliedsonia tecnológicah. contreras-ortiz debolívar |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |