CN111415324B - Classification and identification method for brain disease focus image space distribution characteristics based on magnetic resonance imaging - Google Patents
Classification and identification method for brain disease focus image space distribution characteristics based on magnetic resonance imaging Download PDFInfo
- Publication number
- CN111415324B CN111415324B CN201910737818.4A CN201910737818A CN111415324B CN 111415324 B CN111415324 B CN 111415324B CN 201910737818 A CN201910737818 A CN 201910737818A CN 111415324 B CN111415324 B CN 111415324B
- Authority
- CN
- China
- Prior art keywords
- focus
- brain
- image
- space
- partition
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000002595 magnetic resonance imaging Methods 0.000 title claims abstract description 18
- 208000014644 Brain disease Diseases 0.000 title claims abstract description 8
- 210000004556 brain Anatomy 0.000 claims abstract description 67
- 238000012216 screening Methods 0.000 claims abstract description 14
- 230000011218 segmentation Effects 0.000 claims abstract description 13
- 238000000605 extraction Methods 0.000 claims abstract description 6
- 238000010801 machine learning Methods 0.000 claims abstract description 4
- 238000005192 partition Methods 0.000 claims description 45
- 230000003902 lesion Effects 0.000 claims description 28
- 210000004885 white matter Anatomy 0.000 claims description 11
- 206010051290 Central nervous system lesion Diseases 0.000 claims description 10
- JXSJBGJIGXNWCI-UHFFFAOYSA-N diethyl 2-[(dimethoxyphosphorothioyl)thio]succinate Chemical compound CCOC(=O)CC(SP(=S)(OC)OC)C(=O)OCC JXSJBGJIGXNWCI-UHFFFAOYSA-N 0.000 claims description 6
- 210000001259 mesencephalon Anatomy 0.000 claims description 6
- 206010020649 Hyperkeratosis Diseases 0.000 claims description 5
- 238000012360 testing method Methods 0.000 claims description 5
- 230000009466 transformation Effects 0.000 claims description 5
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 210000001175 cerebrospinal fluid Anatomy 0.000 claims description 4
- 230000000694 effects Effects 0.000 claims description 4
- 210000004884 grey matter Anatomy 0.000 claims description 4
- 238000007477 logistic regression Methods 0.000 claims description 3
- 230000035945 sensitivity Effects 0.000 claims description 3
- 238000013528 artificial neural network Methods 0.000 claims description 2
- 210000005013 brain tissue Anatomy 0.000 claims description 2
- 238000004891 communication Methods 0.000 claims description 2
- 230000008030 elimination Effects 0.000 claims description 2
- 238000003379 elimination reaction Methods 0.000 claims description 2
- 238000010606 normalization Methods 0.000 claims description 2
- 238000002360 preparation method Methods 0.000 claims description 2
- 238000007637 random forest analysis Methods 0.000 claims description 2
- 238000010187 selection method Methods 0.000 claims description 2
- 238000012706 support-vector machine Methods 0.000 claims description 2
- 238000012549 training Methods 0.000 claims description 2
- 238000001514 detection method Methods 0.000 claims 2
- 210000000877 corpus callosum Anatomy 0.000 claims 1
- 238000005259 measurement Methods 0.000 claims 1
- 238000011160 research Methods 0.000 abstract description 6
- 108090000623 proteins and genes Proteins 0.000 abstract description 5
- 238000012545 processing Methods 0.000 abstract description 2
- 238000004458 analytical method Methods 0.000 abstract 2
- 201000010099 disease Diseases 0.000 description 10
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 10
- 108010036280 Aquaporin 4 Chemical group 0.000 description 4
- 102000012002 Aquaporin 4 Human genes 0.000 description 4
- 238000012850 discrimination method Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 208000029067 Neuromyelitis optica spectrum disease Diseases 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 241000257303 Hymenoptera Species 0.000 description 1
- 208000035977 Rare disease Diseases 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000001490 effect on brain Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- 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
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
-
- 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)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention belongs to the technical field of image processing and application, and particularly relates to a classification and identification method of brain disease focus image space distribution characteristics based on magnetic resonance imaging. The method mainly comprises the steps of focus segmentation, individual image registration, space standardization, standard space template individuation, focus space distribution feature extraction, feature screening, modeling and the like, and the core is that a set of analysis method of brain focus image space distribution feature set is constructed through various feature analysis of focuses in individual space and standard space, and machine learning is used for feature screening and modeling on the basis. The method can be used for classifying and identifying brain disease focus images of different brain diseases or brain states caused by different antibodies, different genes and the like by using brain magnetic resonance images, and provides effective guidance for clinical and scientific research.
Description
Technical Field
The invention belongs to the field of image processing and application, and particularly relates to a classification and identification method of brain disease focus image space distribution characteristics based on magnetic resonance imaging.
Background
In the prior art, in the classification and identification method of different brain diseases or brain state images, the brain magnetic resonance imaging technology plays an important role due to the noninvasive property, timeliness and excellent display effect on brain lesions. In clinical practice, doctors often summarize and summarize different characteristics of different disease focuses on images through long-term clinical experience, and perform macroscopic classification identification and report. However, the classification and identification based on experience has the defects of low efficiency, difficulty in finding new characteristics of a focus, difficulty in automatically combining a plurality of characteristics and the like, and the accuracy and the efficiency of classification and identification of the brain focus images are greatly reduced in the face of different diseases, different symptoms or states of genes or antibodies and the like which are similar in clinical manifestation and novel disease subtypes without experience or clinical rare diseases.
The image histology and the deep learning method developed in recent years provide ideas for classifying and identifying brain focus images from the perspective of data analysis. In general, the deep learning method [1] generally requires a large amount of data, and the results thereof often have no interpretability, so that the method is not applicable to small sample sizes and exploration and research phases of similar diseases, rarely seen diseases, new subtypes and other classification problems; the image group learning method [2] carries out quantitative feature extraction including statistical features, texture features, filtering features and the like based on the gray information of the image focus, and then adopts a machine learning method to carry out model construction; the characteristics are manually defined and extracted, so that the result has interpretability; however, traditional image histology often focuses only on features of the lesion itself, not its spatial distribution in the brain. In fact, according to clinical and research experience, diseases caused by different genotypes or antibodies and the like often have important differences in the spatial position and distribution characteristics of the focus of the diseases in the brain, and therefore have important significance in classification and identification.
At present, no systematic classification and identification method for the spatial distribution characteristics of brain lesions based on magnetic resonance imaging exists. Some spatial distribution features have been reported, but are generally identified by the naked eye of a physician or researcher (e.g., whether there are lesions near the ventricle) and then calibrated and generalized counted [3]. On one hand, the method relies on qualitative identification and counting of doctors, the extracted characteristics are limited, and the method is not objective enough, so that differences exist among discrimination results of different doctors or researchers; on the other hand, automation is not realized, and the efficiency is low.
Based on the defects of the existing classification and identification technology in terms of lesion space distribution characteristics, the inventor of the application aims to provide a classification and identification method of space distribution characteristics of brain focus images based on magnetic resonance imaging, and aims to obtain a model and a discrimination method suitable for classification and identification by calculating and analyzing various types of space distribution characteristics of brain focus images based on magnetic resonance imaging, and particularly provides effective guidance for clinical and scientific research in terms of classification of different diseases caused by different genes or antibodies and the like.
The prior art related to the present invention is:
[1]Najafabadi M M，Villanustre F，Khoshgoftaar T M，et al.Deep learning applications and challenges in big data analytics[J].Journal of Big Data，2015，2(1)：1.
[2]Ma X，Zhang L，Huang D，et al.Quantitative radiomic biomarkers for discrimination between neuromyelitis optica spectrum disorder and multiple sclerosis[J].Journal of Magnetic Resonance Imaging，2018.
[3]Jurynczyk M，Geraldes R，Probert F，et al.Distinct brain imaging characteristics of autoantibody-mediated CNS conditions and multiple sclerosis[J].Brain，2017，140(3)：617-627.
disclosure of Invention
The invention aims to provide a classification and identification method of brain focus image space distribution characteristics based on magnetic resonance imaging based on the defects of the existing classification and identification technology in focus space distribution characteristics, which mainly comprises the steps of focus segmentation, individual image registration, space standardization, standard space template individuation, focus space distribution characteristic extraction, characteristic screening, modeling and the like (shown in figure 1). The method of the invention obtains a model and a discrimination method suitable for classification and discrimination by calculating and analyzing various types of spatial distribution characteristics of the magnetic resonance imaging brain focus image, and particularly provides effective guidance for clinic and scientific research in terms of different disease classifications caused by different genes or antibodies and the like.
Specifically, the classification and identification method of the brain focus image space distribution characteristics based on the magnetic resonance imaging comprises the following steps:
1) Segmentation of lesion images:
1) -1, preparation and selection of brain image data: at least two groups of test images G1 and G2; for each tested individual, preparing two images, wherein one image is a focus display image, and selecting one mode from clinical brain magnetic resonance images (including but not limited to T1 weighted images, T2 weighted images, FLAIR, DWI, ADC, SWI and the like) of a patient as an image for subsequent focus segmentation; the second image is a brain structural image, and a T1 weighted image is generally selected; both images may be identical;
1) -2, lesion segmentation: for a "lesion display image", the whole brain lesion is segmented and saved as a binary image or data using image segmentation software, kits, and segmentation algorithms, including but not limited to, MRI vendor provided workstation and segmentation software, MRIcro, MRIcron, MIPAV, ITK-SNAP, MITK, region growing algorithm, etc., to generate a "lesion image";
2) Individual image registration: rigidly registering an individual's "brain structure image" with a "lesion display image", available software and methods include, but are not limited to, SPM, FSL, etc.;
3) Spatial normalization: registering an individual "brain structural image" to a standard space (e.g., MNI standard space, talapiach standard space, etc.), using software and methods including, but not limited to SPM, FSL, AFNI, ANTs, deformation field based registration methods, etc.; applying the transformation parameters generated in the process to the focus image, and obtaining a binarized standard space focus image by selecting a threshold value (default 0.5, optional 0-1); using the same software and method, registering the individual 'brain structural image' with a standard symmetrical brain template (a template with completely symmetrical left and right brain spaces) to obtain a binary 'symmetrical space focus image';
4) Individualizing a standard space template: this step is applicable in the case where in the subsequent steps 5) -4-6, the partition characteristics need to be calculated in the individual space, which may be omitted if performed in the standard space; aiming at partition templates of standard space (including but not limited to gray templates, white matter templates, cerebrospinal fluid templates, brain lobe partition templates, under-curtain structure partition templates, AAL partition templates, brodmann partition templates, harvard Oxford partition templates, white matter partitions such as callus, ventricle templates, self-made partitions such as midbrain aqueduct and the like), extracting each partition in the templates to be an independent binary template, registering all partition templates to the individual space respectively by utilizing the inverse transformation process from the individual to the standard space in the step 3, and obtaining binary individual space partition images corresponding to each partition by selecting a threshold value (default 0.5, optional 0-1);
5) Extracting focal space distribution characteristics: the method comprises the steps of establishing a focus space distribution feature set and an extraction method thereof, wherein the focus space distribution feature set comprises the following categories:
5) -1, distribution size characteristics: extracting single 2D or 3D focus by using individual space focus image according to voxel-to-voxel communication principle, and calculating the size characteristics of single focus such as volume, longest diameter and the like; on the basis of this, the maximum value, average value, sum, median, number of lesions, and large lesion (e.g. more than 200 mm) of the above characteristics of all individual lesions per test are obtained 3 ) As the distribution size feature set of the tested focus;
5) -2, symmetry features: using a symmetrical space focus image and a corresponding partition template, respectively calculating symmetry characteristics of focuses on left and right half brains in the whole brains and each group of symmetrical brain partitions, wherein the symmetry characteristics comprise single side/double side (the left/double/right sides are respectively marked as-1/0/1), differences or ratios of the number of voxels of the focuses on the left and right half brains, a Dice coefficient for measuring distribution similarity and the like; wherein the Dice coefficient of a certain focus i is defined as the double ratio of the number of voxels in the focus intersection and the number of voxels in the union after left-right turning of the left focus (VL) or right focus (VR) images of the symmetrical focus, namely
D(i)＝2×|VL∩VR|/|VL∪VR|
5) -3, probability distribution profile features: for each group tested, calculating the number of non-zero values of each voxel in the group by dividing the number of individuals in the group by the number of all individual binarized 'standard space focus images' in the group to obtain a focus space probability distribution map of the group (the closer the voxel value range is 0-1, the higher the voxel value range is 1, the more the existence probability of the focus of the voxel is represented), and obtaining a high probability distribution map of the group by selecting a threshold value (default 0.2, optional 0-1); defining the characteristic value of the probability distribution map of the focus of a certain tested i in the g group in M groups as the difference value of the sum of the number of voxels in the intersection of the focus area of the i and the high probability distribution focus area of the g group and the number of intersections of the focus and the high probability distribution focus areas of other groups:
the following steps 5) -4-6 can be performed in the individual space by using the focus image and the individual space partition image, or can be performed in the standard space by using the standard space focus image and the partition template of the standard space, or both;
5) -4, single brain partition distribution profile: for each individual brain partition, the distribution characteristics of the brain lesions tested within that partition are calculated: the existence of focus in subarea (1 exists, 0 does not exist), the volume of focus in subarea, the proportion of the volume of focus in subarea to the whole subarea volume, etc.;
5) -5, paracephaly distribution features: performing small region expansion operation (generally not more than 5 mm) on the interested marked brain region (such as ventricle, midbrain aqueduct, callus, white matter region and the like), generating a brain region side region, and calculating the brain region side distribution characteristics (such as ventricle side, midbrain aqueduct periphery, near callus, near white matter region and the like) of the tested focus on the corresponding brain region, including whether the brain region side focus (the existence of 1 and the nonexistence of 0), the brain region side focus volume and the like;
5) -6, distribution of features across brain regions: calculating the cross-brain region characteristics of the tested brain lesion according to the results of the steps 5) -4, wherein the cross-brain region characteristics comprise the number and the two-two ratio of the covered brain tissue types (gray matter/white matter/cerebrospinal fluid), the number of the covered brain regions and the like;
6) Feature screening and modeling:
6) -1, feature screening: on the basis of the focus space distribution feature set extracted in the step 5), adopting various feature screening methods in machine learning to screen features; feature screening methods that can be used include variance selection, chi-square inspection, U-inspection, mutual information, recursive elimination, feature selection methods based on a logistic regression model of L1 penalty term/L2 penalty term/L1 combined with L2 penalty term (e.g., LASSO method), exhaustive methods, and the like; some of these methods (e.g., LASSO methods) can perform feature screening and generate classification discrimination models simultaneously, so that the 6- (2) steps can be combined;
6) -2, establishing a classification discrimination model: taking the screened tested focus characteristics as input, taking the actual tested grouping category as a label, training a classifier, and generating a linear or nonlinear classification identification model; the usable classifier comprises logistic regression, random forest, support vector machine, artificial neural network and the like; after the model is built, the classification and identification effects of the model are evaluated through AUC values and accuracy, sensitivity, specificity values of the ROC curve.
The invention provides a classification and identification method of the spatial distribution characteristics of brain focus images based on magnetic resonance imaging, which is used for obtaining a model and a discrimination method suitable for classification and identification by calculating and analyzing various types of spatial distribution characteristics of brain focus images based on magnetic resonance imaging, and can provide effective guidance for clinic and scientific research especially in the aspects of different diseases classification caused by different genes or antibodies and the like.
Drawings
FIG. 1 is a flow chart of a classification and discrimination method based on spatial distribution characteristics of images of brain lesions of magnetic resonance imaging.
Figure 2 shows a graph of lesion segmentation and extraction results for MOG and AQP4 groups of example 1.
Fig. 3 shows the feature selection result of example 1.
Fig. 4 shows the ROC curve of the model classification discrimination effect of example 1.
Detailed Description
Example 1 differential identification of spatial distribution characteristics of brain foci images in MOG antibody-positive and AQP4 antibody-positive NMOSD patients
Classification and identification:
1) Clinical MRI images of two groups of MOG antibody positive and AQP4 antibody positive NMOSD patients were selected, 28 and 57 images, respectively, each containing FLAIR images as "lesion display images", and T1 weighted images as "brain structural images";
2) The FLAIR image is divided into whole brain lesions by using MRIcron and stored as a binary image as a focus image;
3) Rigidly registering the T1 weighted image of the individual with the FLAIR image using SPM;
4) Registering the individual T1 weighted images to the MNI standard space using SPM; applying the transformation parameters to the FLAIR image, and selecting a threshold value of 0.5 to obtain a binarized standard space focus image;
5) Using the individual space "lesion image", extracting individual 3D lesions (as shown in fig. 2), calculating the volume of individual lesions; obtaining the maximum, average, sum, number of lesions, large lesion (over 200 mm) of the volumes of all individual 3D lesions per test 3 ) Is the number of (3);
6) For each individual brain partition, calculate the presence or absence of the tested brain lesion in that partition (presence of 1, absence of 0), the volume of the lesion in the partition;
7) Calculating the ratio of gray matter and white matter covered by the tested brain focus, and the number of the brain covered partitions;
8) Feature screening and modeling: taking the extracted features in the step 5 as input, taking the actual group category to be tested as a label (MOG group is 1, AQP4 group is 0), and using a LASSO method, wherein 5-fold cross validation is used for feature screening and modeling; the final model contains 9 parameters, of which 1 is constant and the remaining 8 are spatially distributed features (as shown in fig. 3).
The results showed that auc=0.959, accuracy=0.959, sensitivity=1, specificity=0.86 of the ROC curve of the established model had a good classification effect (as shown in fig. 4).
Claims (5)
1. The classifying and identifying method of brain disease focus image space distribution characteristic based on magnetic resonance imaging is characterized by comprising the following steps:
1) Segmentation of lesion images:
2) Individual image registration: carrying out rigid registration on a brain structural image of an individual and a focus display image;
3) Spatial normalization: registering an individual 'brain structural image' to a standard space, applying transformation parameters generated in the process to a 'focus image', and obtaining a binarized 'standard space focus image' by selecting a threshold value; registering the individual brain structural image with a standard symmetrical brain template by using the same software and method to obtain a binarized symmetrical space focus image;
4) Individualizing a standard space template: extracting each partition in the templates as an independent binary template aiming at the partition templates of the standard space, registering all the partition templates to the individual space respectively by utilizing the inverse transformation process from the individual to the standard space in the step 3, and obtaining a binary individual space partition image corresponding to each partition by selecting a fixed number in a 0-1 interval as a threshold value;
5) Extracting focal space distribution characteristics: a focus space distribution feature set of several categories and an extraction method thereof are established:
the method comprises the following steps:
5) -1, distribution size characteristics: extracting single 2D or 3D focus by using an individual space focus image according to the inter-voxel communication principle, and calculating the volume and longest diameter size characteristics of the single focus image; obtaining the maximum value, average value, sum, median, number of lesions, exceeding 200mm of the above features of all single lesion images of each test 3 The number statistical characteristics of the large focus are used as a distribution size characteristic set of the tested focus image;
5) -2, symmetry features: using a symmetrical space focus image and a corresponding partition template, respectively calculating symmetry characteristics of the focus image on the left half brain and the right half brain in the whole brain and each group of symmetrical brain partitions, wherein the symmetry characteristics comprise single side/double side characteristics, the left side/double side/right side characteristics are respectively marked as-1/0/1, and differences or ratios of the number of voxels of the focus image on the left half brain and the right half brain and the difference coefficient of measurement distribution similarity are calculated; wherein the Dice coefficient of a certain focus i is defined as the double ratio of the number of voxels in the focus intersection and the number of voxels in the union after left-right turning of the left focus (VL) or right focus (VR) images of the symmetrical focus, namely
D(i)＝2×|VL∩VR|/|VL∪VR|
5) -3, probability distribution profile features: for each group to be tested, calculating the number of non-zero values of each voxel in the group by using the binarized standard space focus images of all the individuals in the group, dividing the number by the total number of the individuals in the group to obtain a focus space probability distribution map of the group, wherein the closer the voxel value range is 0-1, the higher the existence probability of the focus of the voxel is represented by the closer the voxel value range is, and obtaining a high probability distribution map of the group by selecting a fixed number in a 0-1 interval as a threshold value; defining the characteristic value of the probability distribution map of the focus of a certain tested i in the g group in M groups as the difference value of the sum of the number of voxels in the intersection of the focus area of the i and the high probability distribution focus area of the g group and the number of intersections of the focus and the high probability distribution focus areas of other groups:
steps 5) -4-6 below, analyzing in the individual space using "lesion image" and "individual space partition image", or analyzing in the standard space using "standard space lesion image" and partition template of the standard space, or both;
5) -4, single brain partition distribution profile: for each individual brain partition, the distribution characteristics of the brain lesion images tested within that partition are calculated: the existence of the focus in the subarea is 1, the existence of the focus in the subarea is 0, the volume of the focus in the subarea is not 0, and the proportion of the volume of the focus in the subarea to the whole subarea volume is not zero;
5) -5, paracephaly distribution features: performing a small region expansion operation, no more than 5mm, on a marked brain region of interest including a ventricle, a midbrain aqueduct, a callus, a white matter region, generating a paracerebral region, and calculating a paracerebral distribution characteristic of a tested focus for the corresponding brain region including a paraventricle, a surrounding midbrain aqueduct, a near callus and a near white matter region, including whether a paracerebral focus exists, a presence of 1, a non-presence of 0, and a paracerebral focus volume;
5) -6, distribution of features across brain regions: calculating the cross-brain region distribution characteristics of the tested brain focus image according to the results of the steps 5) -4, wherein the cross-brain region distribution characteristics comprise the covered brain tissue types: the number of grey matter/white matter/cerebrospinal fluid and the number of brain-covering subareas in a two-to-two ratio;
6) Feature screening and modeling.
2. The method of claim 1, wherein in said step 3),
registering an individual "brain structural image" to an MNI standard space or a Talairach standard space; the threshold defaults to a fixed number within the interval of 0-1; the standard symmetrical brain template is a template with completely symmetrical left and right brain spaces.
3. The method of claim 1, wherein said step 4) is adapted to the case where the partition characteristics are to be calculated in the individual space in the subsequent steps 5) -4-6;
the standard space partition templates include, but are not limited to, gray matter templates, white matter templates, cerebrospinal fluid templates, brain lobe partition templates, under-screen structure partition templates, AAL partition templates, brodmann partition templates, harvard Oxford partition templates, corpus callosum white matter partitions, ventricle templates, and midbrain drainage tube homemade partitions.
4. The method of claim 1, wherein said step 1) comprises the sub-steps of:
1) -1, preparation and selection of brain image data: at least two groups of test images G1 and G2; for each tested individual, preparing two images, wherein one image is a focus display image, and selecting one mode from the acquired clinical brain magnetic resonance images of the patient including but not limited to a T1 weighted image, a T2 weighted image and FLAIR, DWI, ADC, SWI as an image for subsequent focus segmentation; the second image is a brain structural image and is selected from T1 weighted images;
1) -2, lesion segmentation: for "lesion display images," the whole brain lesion images are segmented and saved as binary images or data using image segmentation software, kits, and segmentation algorithms, including but not limited to, MRI vendor supplied workstation and segmentation software, MRIcro, MRIcron, MIPAV, ITK-SNAP, MITK, or region growing algorithm, to generate "lesion images".
5. The method of claim 1, wherein said step 6) comprises the substeps of:
5) -1, feature screening: on the basis of the focus image space distribution feature set extracted in the step 5), adopting various feature screening methods in machine learning to screen features; the feature screening method comprises variance selection, chi-square detection, U detection, mutual information, recursion elimination method, feature selection method based on L1 penalty item/L2 penalty item/L1 combined with L2 penalty item or exhaustive method;
5) -2, establishing a classification discrimination model: taking the screened tested focus image characteristics as input, taking the actual tested grouping category as a label, training a classifier, and generating a linear or nonlinear classification identification model; the classifier used comprises logistic regression, random forest, support vector machine or artificial neural network;
after modeling, the classification discrimination effect of the model was evaluated by AUC values, accuracy, sensitivity, specificity values of ROC curves.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910737818.4A CN111415324B (en) | 2019-08-09 | 2019-08-09 | Classification and identification method for brain disease focus image space distribution characteristics based on magnetic resonance imaging |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910737818.4A CN111415324B (en) | 2019-08-09 | 2019-08-09 | Classification and identification method for brain disease focus image space distribution characteristics based on magnetic resonance imaging |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111415324A CN111415324A (en) | 2020-07-14 |
| CN111415324B true CN111415324B (en) | 2024-03-08 |
Family
ID=71494186
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910737818.4A Active CN111415324B (en) | 2019-08-09 | 2019-08-09 | Classification and identification method for brain disease focus image space distribution characteristics based on magnetic resonance imaging |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111415324B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112651976B (en) * | 2020-10-13 | 2022-07-19 | 首都医科大学附属北京天坛医院 | Focal region brain network determination method and system based on low-resolution nuclear magnetic data |
| CN112075927B (en) * | 2020-10-15 | 2024-05-14 | 首都医科大学附属北京天坛医院 | Etiology classification method and device for cerebral apoplexy |
| CN113283465B (en) * | 2021-04-02 | 2022-04-29 | 电子科技大学 | Diffusion tensor imaging data analysis method and device |
| CN113192014B (en) * | 2021-04-16 | 2024-01-30 | 深圳市第二人民医院(深圳市转化医学研究院) | Training method and device for improving ventricle segmentation model, electronic equipment and medium |
| CN113066113B (en) * | 2021-04-27 | 2022-01-11 | 河北医科大学第二医院 | Method and equipment for constructing spontaneous hypertensive rat brain template and map set |
| CN118097245A (en) * | 2024-02-20 | 2024-05-28 | 深圳市儿童医院 | Brain nodule load generation method based on MRI and related equipment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106156484A (en) * | 2016-06-08 | 2016-11-23 | 中国科学院自动化研究所 | Disease of brain individuation Forecasting Methodology based on nuclear magnetic resonance image and system |
| WO2017210873A1 (en) * | 2016-06-08 | 2017-12-14 | 中国科学院自动化研究所 | Individualized brain disease prediction method and system based on magnetic resonance imaging |
-
2019
- 2019-08-09 CN CN201910737818.4A patent/CN111415324B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106156484A (en) * | 2016-06-08 | 2016-11-23 | 中国科学院自动化研究所 | Disease of brain individuation Forecasting Methodology based on nuclear magnetic resonance image and system |
| WO2017210873A1 (en) * | 2016-06-08 | 2017-12-14 | 中国科学院自动化研究所 | Individualized brain disease prediction method and system based on magnetic resonance imaging |
Non-Patent Citations (2)
| Title |
|---|
| 基于VBM-DARTEL的局灶性皮质发育不良脑灰质体积异常检测方法;庞清华等;《中国医学物理学杂志》;20160425(第04期);全文 * |
| 纹理分析在甲状腺结节影像学中的研究进展;张衡等;《中国医学计算机成像杂志》;20181025(第05期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111415324A (en) | 2020-07-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111415324B (en) | Classification and identification method for brain disease focus image space distribution characteristics based on magnetic resonance imaging | |
| Yoo et al. | Prostate cancer detection using deep convolutional neural networks | |
| CN111488914B (en) | Alzheimer disease classification and prediction system based on multitask learning | |
| CN111599464B (en) | Novel multi-mode fusion auxiliary diagnosis method based on rectal cancer image histology | |
| CN112101451B (en) | Breast cancer tissue pathological type classification method based on generation of antagonism network screening image block | |
| Bjornsson et al. | Associative image analysis: a method for automated quantification of 3D multi-parameter images of brain tissue | |
| US8060348B2 (en) | Systems for analyzing tissue samples | |
| US8831327B2 (en) | Systems and methods for tissue classification using attributes of a biomarker enhanced tissue network (BETN) | |
| Ahirwar | Study of techniques used for medical image segmentation and computation of statistical test for region classification of brain MRI | |
| Alksas et al. | A novel computer-aided diagnostic system for accurate detection and grading of liver tumors | |
| CN109840554B (en) | Alzheimer's disease MRI image classification method based on SVM-RFE-MRMR algorithm | |
| CN102855491B (en) | A kind of central brain functional magnetic resonance image classification Network Based | |
| CN113616184A (en) | Brain network modeling and individual prediction method based on multi-mode magnetic resonance image | |
| CN112418337B (en) | Multi-feature fusion data classification method based on brain function hyper-network model | |
| US10423820B2 (en) | Systems and methods for automatic generation of training sets for machine interpretation of images | |
| CN103020653B (en) | Structure and function magnetic resonance image united classification method based on network analysis | |
| CN102509123A (en) | Brain functional magnetic resonance image classification method based on complex network | |
| CN108122221A (en) | The dividing method and device of diffusion-weighted imaging image midbrain ischemic area | |
| CN113610808A (en) | Individual brain atlas individualization method, system and equipment based on individual brain connection atlas | |
| CN114600155A (en) | Weakly supervised multitask learning for cell detection and segmentation | |
| CN101699515B (en) | Multi-elite immune quantum clustering-based medical image segmenting system and multi-elite immune quantum clustering-based medical image segmenting method | |
| Wang et al. | Application of neuroanatomical features to tractography clustering | |
| CN116188423A (en) | Super-pixel sparse and unmixed detection method based on pathological section hyperspectral image | |
| CN115546605A (en) | Training method and device based on image labeling and segmentation model | |
| CN113936172A (en) | Disease classification method and device based on ensemble learning and multi-mode feature fusion |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |