WO2018129650A1

WO2018129650A1 - Analysis method for multi-mode radiomics, apparatus and terminal

Info

Publication number: WO2018129650A1
Application number: PCT/CN2017/070740
Authority: WO
Inventors: 宋柏霖; 李志成; 李其花; 孙秋畅; 王梦巧
Original assignee: 中国科学院深圳先进技术研究院
Priority date: 2017-01-10
Filing date: 2017-01-10
Publication date: 2018-07-19

Abstract

An analysis method for multi-mode radiomics, an apparatus and a terminal, said method comprising: acquiring a plurality of modal images, and pre-processing said plurality of modal images (S101); performing region segmentation on a pre-processed modal image, and obtaining a region of interest which corresponds to each modal image (S102); performing high-throughput characteristic extraction on each region of interest of each modal image, and obtaining a characteristic which corresponds to each region of interest (S103); forming a source characteristic by using the characteristic which corresponds to each region of interest of said plurality of modal images, and using a preset clustering algorithm to perform characteristic clustering on said source characteristic (S104); constructing a radiomics marker according to a result of characteristic clustering (S105), thereby solving the problem in the prior art wherein image characteristics cannot be extracted from multiple aspects during radiomics research, and increasing the number of source characteristics used for characteristic clustering and representative characteristic categories following clustering, thus mining medical image information to the greatest extent.

Description

Analysis method, device and terminal for multimodal image omics

Technical field

The invention belongs to the technical field of biomedical engineering, and in particular relates to a method, a device and a terminal for analyzing multi-modal image omics.

Background technique

Traditional medical image analysis generally focuses only on the manifestation of tumors in a single modality. Image omics can transform traditional medical images into digable data. In the imaging omics research, the prior art mainly acquires the static information of the four anatomical images without considering the dynamic information contained in the image, and cannot extract the image features in many aspects, thereby resulting in the extracted feature numbers and sample cases. Limited, can not maximize the mining of medical image information.

Summary of the invention

In view of this, an embodiment of the present invention provides a method, a device, and a terminal for analyzing multi-modal image omics to solve the problem that the prior art cannot extract image features in multiple aspects in image omics research, so as to achieve maximum Mining medical image information.

In a first aspect, an analysis method for multimodal image omics is provided, the method comprising:

Obtaining a plurality of modal images and preprocessing the plurality of modal images;

Performing region segmentation on the modal image after preprocessing to obtain a region of interest corresponding to each modal image;

Perform high-throughput feature extraction for each region of interest of each modal image to acquire features corresponding to each region of interest;

Generating the source features by the features corresponding to each of the regions of interest of the plurality of modal images, using pre- a clustering algorithm is used to perform feature clustering on the source features;

Image omics markers were constructed based on the results of feature clustering.

Further, the plurality of modal images includes four MR anatomical imaging and diffusion tensor imaging, diffusion weighted imaging, and dynamic contrast enhanced imaging;

The four MR anatomical images include T1 weighted imaging, T1 contrast enhanced imaging, T2 weighted imaging, T2 flow attenuation inversion recovery sequence imaging.

Further, the acquiring a plurality of modal images and performing preprocessing on the plurality of modal images includes:

Obtain multiple modal images;

Image registration, smoothing, and interpolation processing are performed on the plurality of modal images.

Further, performing image registration on the plurality of modal images includes:

T1 contrast enhanced imaging in four MR anatomical images was selected as the reference image modal;

Obtaining a spatial coordinate transformation parameter by a similarity measure;

And storing the remaining modal images in the plurality of modal images with the T1 contrast enhanced imaging according to the spatial coordinate transformation parameter.

Further, the features corresponding to the region of interest include morphological features, grayscale features, and texture features.

In a second aspect, an analysis apparatus for multimodal image omics is provided, the apparatus comprising:

a preprocessing module, configured to acquire a plurality of modal images, and preprocess the plurality of modal images;

a segmentation module, configured to perform region segmentation on the modal image after preprocessing, and obtain a region of interest corresponding to each modal image;

a feature extraction module, configured to perform high-throughput feature extraction for each region of interest of each modal image, and acquire features corresponding to each region of interest;

a feature clustering module, configured to form a source feature by using a feature corresponding to each region of the plurality of modal images, and performing feature clustering on the source feature by using a preset clustering algorithm;

A building block is configured to construct an ombryographic marker based on the results of feature clustering.

Further, the preprocessing module includes:

An acquisition unit for acquiring a plurality of modal images;

And a processing unit, configured to perform image registration, smoothing, and interpolation processing on the plurality of modal images.

Further, the processing unit is specifically configured to:

In a third aspect, a terminal is provided, the terminal comprising a processor, the processor for executing the following program modules of the presence memory:

Compared with the prior art, the embodiment of the present invention obtains a plurality of modal images and performs pre-processing on the plurality of modal images; and then performs segmentation on the pre-processed modal images to obtain each a region of interest corresponding to the modal image; performing high-throughput feature extraction for each region of interest of each modal image to acquire features corresponding to each region of interest; and finally each of the plurality of modal images A feature corresponding to a region of interest constitutes a source feature, and the source feature is clustered by a preset clustering algorithm, and an image ensemble marker is constructed according to the result of the feature clustering; thereby solving the prior art image In the omics research, the problem of image features cannot be extracted in many aspects, which greatly enriches the number of source features used for feature clustering and the representative feature types after feature clustering, and realizes the maximum mining of medical image information.

DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below. Obviously, the drawings in the following description are only It is a certain embodiment of the present invention, and other drawings can be obtained from those skilled in the art without any inventive effort.

1 is a flowchart of an implementation of an analysis method for multi-modal image omics provided by an embodiment of the present invention;

2 is a specific implementation process of step S101 in the analysis method of multi-modal image omics provided by the embodiment of the present invention;

3 is a schematic flowchart showing an implementation process of clustering analysis of glioma by K-means clustering algorithm according to an embodiment of the present invention;

FIG. 4 is a structural diagram of a multi-modal image omics analysis apparatus according to an embodiment of the present invention.

detailed description

The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

In the embodiment of the present invention, a plurality of modal images are acquired, and the plurality of modal images are preprocessed; and then the modal images after preprocessing are segmented to obtain a corresponding sensation of each modal image. Interesting region; performing high-throughput feature extraction for each region of interest of each modal image to acquire features corresponding to each region of interest; and finally corresponding to each region of interest of the plurality of modal images The feature is composed of source features, and the source features are clustered by using a preset clustering algorithm, and the ensemble markers are constructed according to the results of feature clustering; thereby solving the problem that the prior art cannot be used in image group research. The problem of extracting image features greatly enriches the number of source features used for feature clustering and the representative feature types after feature clustering, and achieves maximum mining of medical image information. The embodiments of the present invention also provide corresponding devices, which are described in detail below.

FIG. 1 is a flowchart showing an implementation process of an analysis method for multi-modal image omics provided by an embodiment of the present invention.

In the embodiment of the present invention, the multi-modal image omics analysis method is applied to a computer, a server, and the like.

Referring to FIG. 1, the analysis method of the multimodal image omics includes:

In step S101, a plurality of modal images are acquired, and the plurality of modal images are preprocessed.

Here, the embodiment of the present invention introduces diffusion tensor imaging, diffusion weighted imaging, and dynamic contrast enhanced imaging based on the four MR anatomical imaging modalities used in the original imaging omics research. Therefore, the plurality of modal images described in the embodiments of the present invention include four MR anatomical imaging and diffusion tensor imaging, diffusion weighted imaging, and dynamic contrast enhanced imaging. The four MR anatomical images include T1 weighted imaging, T1 contrast enhanced imaging, T2 weighted imaging, and T2 flow attenuation inversion recovery sequence imaging. The diffusion tensor imaging and the diffusion weighted imaging are imaging modes that display brain complement features according to the motion characteristics of water molecules. The dynamic contrast enhanced imaging can reflect the blood flow dynamics and perfusion status inside the tumor.

In the embodiment of the present invention, different modal images are obtained by different imaging modes through different devices. Due to the different parameters of the image acquisition, the four MR anatomical imaging as well as the diffusion tensor imaging, the diffusion weighted imaging, and the dynamic contrast enhanced imaging need to be pre-processed. Illustratively, FIG. 2 shows a specific implementation flow of step S101 in the analysis method of multi-modal image omics provided by the embodiment of the present invention.

Referring to FIG. 2, the step S101 includes:

In step S201, a plurality of modal images are acquired.

The plurality of modal images, namely the above four MR anatomical imaging and diffusion tensor imaging, diffusion weighting Imaging, dynamic contrast enhanced imaging, each modal image comprising the same number of images, preferably 20 to 40 sheets.

In step S202, image registration, smoothing processing, and interpolation processing are performed on the plurality of modal images.

In the embodiment of the present invention, the preprocessing includes image registration, image smoothing, and interpolation processing. Here, image registration refers to one or a series of spatial transformations of an image to achieve spatial consistency with corresponding pixels on another image. Among them, the way of image registration includes relative registration and absolute registration. Relative registration refers to selecting one image in multiple images as the reference image, and registering other related images, wherein the registration coordinate relationship is arbitrary; absolute registration refers to pre-defining a control grid, which will All images are registered with respect to this grid, ie the corresponding pixels of the respective images are spatially identical by geometrically correcting each image separately.

As a preferred example of the present invention, T1 contrast enhanced imaging in the four MR anatomical images may be selected as a reference image modality; then spatial coordinate transformation parameters are acquired by a similarity metric; and parameters are transformed according to the spatial coordinate transformation parameters The remaining modal images in the plurality of modal images are respectively registered with the T1 contrast enhanced imaging. Specifically, after selecting T1 contrast enhanced imaging as the reference image modality, several images included in the T1 contrast enhanced imaging are used as reference images, and each of the remaining six modal images is contrast enhanced with the T1. The corresponding image in the imaging is subjected to a unified coordinate system conversion.

As a preferred example of the present invention, the smoothing process includes an averaging filter, a median filter, etc., and preferably a median filter is used to perform image smoothing processing on each modal image. In the embodiment of the present invention, the smoothing process is used to filter out the unsmooth burrs and sharp edges introduced during the data acquisition and morphological processing, thereby further ensuring the purity of the modal image after the preprocessing.

It should be noted that the sequence of image registration, image smoothing, and interpolation processing according to the embodiments of the present invention is arranged according to actual requirements; image registration, smoothing, and interpolation may be performed on the modal image first. Processing; or, the modal image is first smoothed and interpolated, and then image registration is performed. The specific order is determined according to the type of features extracted. For example, when extracting grayscale features, shape features, and most texture features, it is preferable to perform image registration and re-entry. Line smoothing and interpolation processing; when extracting a small number of specified texture features, it is preferable to perform smoothing and interpolation processing first, and then perform image registration to improve the feature extraction effect.

In step S102, the modal image after the pre-processing is segmented to obtain a region of interest corresponding to each modal image.

Here, after segmentation, each modal image corresponds to one or more regions of interest. Illustratively, taking the influence analysis of glioma as an example, the T2 flow attenuation inversion recovery image segmentation tumor edema area, that is, the entire tumor area; T1 contrast enhanced imaging segmentation tumor enhancement zone, necrotic zone; diffusion tensor Imaging, diffusion-weighted imaging, and dynamic contrast-enhanced imaging respectively segment the tumor enhancement zone to extract dynamic information such as blood flow information and material distribution; the T1-weighted imaging and T2-weighted imaging are used for comparison of tumor segmentation results.

In step S103, high-throughput feature extraction is performed on each region of interest of each modal image to acquire features corresponding to each region of interest.

In the embodiment of the present invention, one or more regions of interest corresponding to each modal image are obtained by region segmentation, and then a corresponding set of features is extracted by using high-throughput features. The extracted features include morphological features, grayscale features, and texture features. The morphological features are used to describe the three dimensional features of the tumor. The grayscale features are used to describe grayscale values corresponding to all pixels in each region of interest. The texture features are used to quantify heterogeneity within the tumor. Illustratively, Table 1 gives an example of the composition of a set of features corresponding to each extracted region of interest provided by an embodiment of the present invention. In Table 1, the feature set includes 28 morphological features, 12 grayscale features, and 52 texture features.

Table 1

Four MR anatomical imaging, diffusion tensor imaging, diffusion weighted imaging, dynamic pairing Compared with each region of interest of each modal image in the enhanced imaging, a corresponding set of morphological features, grayscale features, and texture features are extracted, which increases image dynamic information and functional change information, such as blood flow, The distribution of matter, thus maximizing the information in different modal images, greatly enriches the number of source features used for feature clustering.

In step S104, the source features are composed of features corresponding to each of the plurality of modal images, and the source features are clustered by using a preset clustering algorithm.

Here, the preset clustering algorithm includes a hierarchical clustering algorithm, a density and grid based clustering algorithm, a K-means clustering algorithm, and the like.

Illustratively, the clustering will be described below by taking the influence analysis of the glioma described above as an example. In order to achieve feature dimension reduction, the typical method K-means algorithm in the partitioned clustering algorithm is used to perform feature clustering. Among them, the purpose of the K-means clustering algorithm is to minimize the sum of the squares of the differences between each source feature and the cluster center. FIG. 3 is a flowchart showing an implementation process of clustering analysis of glioma by K-means clustering algorithm according to an embodiment of the present invention, including:

In step S301, k cluster centers are randomly selected.

In step S302, a source feature is composed of features corresponding to each region of interest of the plurality of modal images, a distance value between each source feature and each cluster center is calculated, and the source feature is assigned to In the class indicated by the cluster center with the smallest value.

In step S303, after the allocation is completed, a deviation value is calculated, the deviation value being a sum of squares of distances between each source feature and the k cluster centers.

In step S304, it is determined whether the deviation value converges.

If yes, step S305 is performed; otherwise, step S301 is returned to re-select k cluster centers for the next round of feature clustering operations.

In step S305, the current clustering operation is ended.

Here, it is assumed that the calculation formula of the deviation value is C ₁ , . . . , C _k by the cluster center selected in step S301:

In the above formula, x _i represents the i-th source feature, and d ² (x _i , C _r ) represents the square of the difference between the i-th source feature and the r-th cluster center, that is, the square of the distance, and D represents The deviation value is used to measure the effect of the K-means algorithm. The smaller the deviation value D is, the better the effect is.

It is found that the number of source features for feature clustering obtained by the embodiment of the present invention can reach 1,564, and the representative feature types after clustering can reach 10 categories, far exceeding the existing literature and patents. Therefore, with the embodiment of the present invention, the number of source features before feature clustering and the representative feature types after feature clustering can be greatly enriched.

In step S105, an ensemble marker is constructed based on the result of feature clustering.

In the embodiment of the present invention, the constructing the image omics marker is to predict and analyze the clinical disease by using a computer automatic identification and classification method. Specifically, all the features of all patients who have completed feature extraction are compared with the parameters of the patient's clinical pathology, survival period, etc., which need to be predicted and analyzed, and the characteristics of different modal images are subjected to "training data set" and "verification data". The classification of the set. The computer automatic identification method is used to train the different types of lesion data in the training data set, and a complete training model is used to verify the data set to realize the analysis and prediction of unknown diseases, and the pathology, clinical stage and genetic information of the patient are obtained. And qualitative and quantitative analysis of survival.

In summary, the embodiment of the present invention obtains a plurality of modal images and performs preprocessing on the plurality of modal images; and then performs region segmentation on the preprocessed modal images to acquire each modal image. Corresponding regions of interest; performing high-throughput feature extraction for each region of interest of each modal image to acquire features corresponding to each region of interest; and finally interested in each of the plurality of modal images The feature corresponding to the region constitutes the source feature, and the source feature is clustered by using a preset clustering algorithm, and the image omics marker is constructed according to the result of the feature clustering; thereby solving the prior art research on image omics The problem of extracting image features in many aspects is not rich, and the number of source features used for feature clustering and the representative feature types after feature clustering are enriched, and medical image information is maximized.

FIG. 4 shows a composition of an apparatus for analyzing multi-modal image omics provided by an embodiment of the present invention. For convenience of description, only parts related to the embodiment of the present invention are shown.

In the embodiment of the present invention, the device is used to implement the multi-modal image omics analysis method described in the foregoing embodiments of FIG. 1 to FIG. 3, and may be a software unit or a hardware unit built in a computer or a server. A unit that combines hardware and software.

Referring to Figure 4, the apparatus includes:

The preprocessing module 41 is configured to acquire a plurality of modal images and perform preprocessing on the plurality of modal images;

The segmentation module 42 is configured to perform region segmentation on the modal image after preprocessing, and acquire a region of interest corresponding to each modal image;

The feature extraction module 43 is configured to perform high-throughput feature extraction for each region of interest of each modal image to acquire features corresponding to each region of interest;

The feature clustering module 44 is configured to form a source feature by using a feature corresponding to each region of the plurality of modal images, and perform feature clustering on the source feature by using a preset clustering algorithm;

The building module 45 is configured to construct an ombryographic marker according to the result of the feature clustering.

In an embodiment of the invention, the plurality of modal images include four MR anatomical imaging and diffusion tensor imaging, diffusion weighted imaging, dynamic contrast enhanced imaging; the four MR anatomical imaging includes T1 weighted imaging, T1 contrast enhancement Imaging, T2-weighted imaging, T2 flow attenuation inversion recovery sequence imaging. The diffusion tensor imaging and the diffusion weighted imaging are imaging modes that display brain complement features according to the motion characteristics of water molecules. The dynamic contrast enhanced imaging can reflect the blood flow dynamics and perfusion status inside the tumor. Each modal image includes the same number of images, preferably 20 to 40 sheets.

Due to the different parameters of the image acquisition, in order to ensure the consistency of the image, the four MR anatomical imaging as well as the diffusion tensor imaging, the diffusion weighted imaging, and the dynamic contrast enhanced imaging are required to be preprocessed. In the embodiment of the present invention, the pre-processing module 41 includes:

The acquiring unit 411 is configured to acquire a plurality of modal images;

The processing unit 412 is configured to perform image registration, smoothing, and interpolation processing on the plurality of modal images.

Here, image registration refers to one or a series of spatial transformations on an image. It is spatially consistent with the corresponding pixel on the other image. Among them, the way of image registration includes relative registration and absolute registration. Relative registration refers to selecting one image in multiple images as the reference image, and registering other related images, wherein the registration coordinate relationship is arbitrary; absolute registration refers to pre-defining a control grid, which will All images are registered with respect to this grid, ie the corresponding pixels of the respective images are spatially identical by geometrically correcting each image separately.

As a preferred example of the present invention, image registration can be performed based on T1 contrast enhanced imaging. In this case, the processing unit 412 can also be used to:

Specifically, after T1 contrast enhanced imaging is selected as the reference image modality, a plurality of images included in the T1 contrast enhanced imaging are used as reference images, and the processing unit 412 compares each of the remaining six modal images with The T1 contrasts the corresponding image in the enhanced imaging to perform a unified coordinate system conversion.

As a preferred example of the present invention, the processing unit 412 may perform smoothing processing using an averaging filter and a median filter, and preferably perform image smoothing processing on each modal image using a median filter. In the embodiment of the present invention, the unsmooth burrs and the sharp edges introduced by the human in the process of data acquisition and morphological processing are filtered out by the smoothing process, thereby further ensuring the purity of the modal image after the preprocessing.

It should be noted that the sequence of image registration, image smoothing, and interpolation processing according to the embodiment of the present invention is arranged according to actual requirements; the pre-processing unit 412 may perform image registration on the modal image first. Then, the smoothing process and the interpolation process are performed; or the pre-processing unit 412 may perform smoothing processing and interpolation processing on the modal image, and then perform image registration. The specific order is determined according to the type of features extracted. For example, when extracting grayscale features, shape features, and most texture features, it is preferred to perform image registration, smoothing, and interpolation; When a small number of specified texture features are selected, it is preferable to perform smoothing processing and interpolation processing, and then perform image registration to improve the feature extraction effect.

Further, in the embodiment of the present invention, the segmentation module 42 performs region segmentation on each modal image to obtain one or more regions of interest corresponding to each modal image, and then performs Qualcomm through the feature extraction module 43. Quantity feature extraction. The extracted features include morphological features, grayscale features, and texture features. The morphological features are used to describe the three dimensional features of the tumor. The grayscale features are used to describe grayscale values corresponding to all pixels in each region of interest. The texture features are used to quantify heterogeneity within the tumor. Table 1 gives an example of the composition of the extracted features provided by the embodiments of the present invention. In Table 1, the features include 28 morphological features, 12 grayscale features, and 52 texture features.

The embodiment of the present invention extracts a corresponding set of morphological features and grays for each region of interest of each of the four MR anatomical imaging, diffusion tensor imaging, diffusion weighted imaging, and dynamic contrast enhanced imaging. Degree features and texture features increase image dynamic information and functional change information, such as blood flow and material distribution, compared to the features extracted by the prior art, thereby greatly enriching the number of source features for feature clustering.

Further, the preset clustering algorithm used by the feature clustering module 44 includes a hierarchical clustering algorithm, a density and grid based clustering algorithm, a K-means clustering algorithm, and the like.

Illustratively, clustering will be described below by taking the analysis of the influence of glioma as an example. In order to achieve feature dimension reduction, the feature extraction module 43 uses the typical method K-means algorithm in the partitioned clustering algorithm to perform feature clustering. Among them, the purpose of the K-means clustering algorithm is to minimize the sum of the squares of the differences between each feature and the cluster center. The feature extraction module 43 is specifically used for.

Randomly select k cluster centers;

Calculating a source feature of each of the plurality of modal images corresponding to each region of interest, calculating a distance value between each source feature and each cluster center, and assigning the source feature to a cluster having the smallest distance value In the class indicated by the class center;

After the allocation is completed, a deviation value is calculated, the deviation value being a sum of squares of distances between each source feature and the k cluster centers;

Determining whether the deviation value converges;

If it is, end this clustering operation. Otherwise, k cluster centers are re-selected for the next round of feature clustering operations.

Here, assuming that the selected cluster centers are C ₁ ,..., C _k , the calculation formula of the deviation value is:

It should be noted that the device in the embodiment of the present invention may be used to implement all the technical solutions in the foregoing method embodiments, and the functions of the respective functional modules may be specifically implemented according to the method in the foregoing method embodiment, and the specific implementation process may refer to The related descriptions in the above examples are not described herein again.

The present invention also provides a terminal for implementing an analysis method for multimodal image omics. In this embodiment, the terminal includes: a processor, wherein the processor is configured to execute the following program modules of the presence memory:

Wherein, the plurality of modal images comprise four MR anatomical imaging and diffusion tensor imaging, diffusion weighted imaging, dynamic contrast enhanced imaging; and the four MR anatomical imaging comprises T1 weighted imaging, T1 contrast enhanced imaging, T2 weighted imaging , T2 flow attenuation inversion recovery sequence imaging.

Further, the preprocessing module includes:

An acquisition unit for acquiring a plurality of modal images;

Further, the processing unit is specifically configured to:

The features corresponding to the region of interest include morphological features, grayscale features, and texture features.

It should be understood that, in the embodiment of the present invention, the so-called processor may be a central processing unit (CPU) and/or a graphics processing unit (GPU), and may also be combined with other general processing on this basis. , Digital Signal Processor (DSP), Application Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic device, discrete door or Transistor logic devices, discrete hardware components, etc.

Optionally, the terminal may further include one or more input devices, one or more output devices. The above processor, input device, output device, and memory are connected by a bus.

The input device may include a touchpad, a fingerprint sensor (for collecting fingerprint information of the user and direction information of the fingerprint), a microphone, a communication module (such as a Wi-Fi module, a 2G/3G/4G network module), and a physical button. Wait.

The output device may include a display (LCD or the like), a speaker, and the like. Among them, the display can be used to display information input by the user or information provided to the user, and the like. The display can include a display panel, optional, The display panel can be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), or the like. Further, the touch screen may be overlaid on the display, and when the touch screen detects a touch operation on or near the touch screen, the touch screen is transmitted to the processor to determine the type of the touch event, and then the processor provides a corresponding display on the display according to the type of the touch event. Visual output.

In a specific implementation, the processor, the input device, the output device, and the memory described in the embodiments of the present invention may implement the implementation manner described in the embodiment of the multi-modal image omics analysis method provided by the embodiment of the present invention. This will not be repeated here.

Those of ordinary skill in the art will appreciate that the elements and algorithm steps of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the solution. A person skilled in the art can use different methods for implementing the described functions for each particular application, but such implementation should not be considered to be beyond the scope of the present invention.

A person skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working process of the device and the unit described above can refer to the corresponding process in the foregoing method embodiment, and details are not described herein again.

In the several embodiments provided by the present application, it should be understood that the disclosed method, apparatus and terminal End, can be achieved in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules and units is only a logical function division. In actual implementation, there may be another division manner, for example, multiple units or components may be used. Combinations can be integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be in an electrical, mechanical or other form.

The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, each functional unit and module in various embodiments of the present invention may be integrated into one processing unit, or each unit or module may exist physically separately, or two or more units or modules may be integrated into one unit. .

The functions may be stored in a computer readable storage medium if implemented in the form of a software functional unit and sold or used as a standalone product. Based on such understanding, the technical solution of the present invention, which is essential or contributes to the prior art, or a part of the technical solution, may be embodied in the form of a software product, which is stored in a storage medium, including The instructions are used to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and the like. .

The above is only a specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily think of changes or substitutions within the technical scope of the present invention. It should be covered by the scope of the present invention. Therefore, the scope of the invention should be determined by the scope of the claims.

Claims

A method for analyzing multimodal image omics, characterized in that the method comprises:

Obtaining a plurality of modal images and preprocessing the plurality of modal images;

Performing region segmentation on the modal image after preprocessing to obtain a region of interest corresponding to each modal image;

Perform high-throughput feature extraction for each region of interest of each modal image to acquire features corresponding to each region of interest;

The source features are composed of features corresponding to each of the plurality of modal images, and the source features are clustered by using a preset clustering algorithm;

Image omics markers were constructed based on the results of feature clustering.
The multimodal image omics analysis method according to claim 1, wherein the plurality of modal images comprise four MR anatomical imaging and diffusion tensor imaging, diffusion weighted imaging, and dynamic contrast enhanced imaging;

The four MR anatomical images include T1 weighted imaging, T1 contrast enhanced imaging, T2 weighted imaging, T2 flow attenuation inversion recovery sequence imaging.
The method for analyzing multimodal image omics according to claim 1 or 2, wherein the acquiring a plurality of modal images and preprocessing the plurality of modal images comprises:

Obtain multiple modal images;

Image registration, smoothing, and interpolation processing are performed on the plurality of modal images.
The method of analyzing multi-modal image omics according to claim 3, wherein performing image registration on the plurality of modal images comprises:

T1 contrast enhanced imaging in four MR anatomical images was selected as the reference image modal;

Obtaining a spatial coordinate transformation parameter by a similarity measure;

And storing the remaining modal images in the plurality of modal images with the T1 contrast enhanced imaging according to the spatial coordinate transformation parameter.
The method for analyzing multimodal image omics according to any one of claims 1, 2, and 4, characterized in that The features corresponding to the region of interest include morphological features, grayscale features, and texture features.
An apparatus for analyzing multimodal image omics, characterized in that the apparatus comprises:

a preprocessing module, configured to acquire a plurality of modal images, and preprocess the plurality of modal images;

a segmentation module, configured to perform region segmentation on the modal image after preprocessing, and obtain a region of interest corresponding to each modal image;

a feature extraction module, configured to perform high-throughput feature extraction for each region of interest of each modal image, and acquire features corresponding to each region of interest;

a feature clustering module, configured to form a source feature by using a feature corresponding to each region of the plurality of modal images, and performing feature clustering on the source feature by using a preset clustering algorithm;

A building block is configured to construct an ombryographic marker based on the results of feature clustering.
The apparatus for analyzing multimodal image omniscience according to claim 6, wherein the plurality of modal images comprise four MR anatomical imaging and diffusion tensor imaging, diffusion weighted imaging, and dynamic contrast enhanced imaging;

The four MR anatomical images include T1 weighted imaging, T1 contrast enhanced imaging, T2 weighted imaging, T2 flow attenuation inversion recovery sequence imaging.
The apparatus for analyzing multimodal image omics according to claim 6 or 7, wherein the preprocessing module comprises:

An acquisition unit for acquiring a plurality of modal images;

And a processing unit, configured to perform image registration, smoothing, and interpolation processing on the plurality of modal images.
The apparatus for analyzing multimodal image omics according to claim 8, wherein the processing unit is specifically configured to:

T1 contrast enhanced imaging in four MR anatomical images was selected as the reference image modal;

Obtaining a spatial coordinate transformation parameter by a similarity measure;

And storing the remaining modal images in the plurality of modal images with the T1 contrast enhanced imaging according to the spatial coordinate transformation parameter.
The apparatus for analyzing multimodal image omics according to any one of claims 6, 7, and 9, The feature corresponding to the region of interest includes morphological features, grayscale features, and texture features.
A terminal, characterized in that the terminal comprises a processor for executing the following program modules of the presence memory:

a preprocessing module, configured to acquire a plurality of modal images, and preprocess the plurality of modal images;

a segmentation module, configured to perform region segmentation on the modal image after preprocessing, and obtain a region of interest corresponding to each modal image;

a feature extraction module, configured to perform high-throughput feature extraction for each region of interest of each modal image, and acquire features corresponding to each region of interest;

a feature clustering module, configured to form a source feature by using a feature corresponding to each region of the plurality of modal images, and performing feature clustering on the source feature by using a preset clustering algorithm;

A building block is configured to construct an ombryographic marker based on the results of feature clustering.