CN113628207B

CN113628207B - Image area segmentation method, device, equipment and storage medium

Info

Publication number: CN113628207B
Application number: CN202111001199.6A
Authority: CN
Inventors: 史张; 林江; 曾蒙苏; 吕鹏; 张冉颖
Original assignee: Naoxi Suzhou Intelligent Technology Co ltd
Current assignee: Naoxi Suzhou Intelligent Technology Co ltd
Priority date: 2021-08-30
Filing date: 2021-08-30
Publication date: 2023-04-07
Anticipated expiration: 2041-08-30
Also published as: CN113628207A

Abstract

The invention provides an image region segmentation method, an image region segmentation device, image region segmentation equipment and a storage medium, wherein the method comprises the following steps: acquiring a cerebral blood flow image and an apparent diffusion coefficient image of a brain of a target object; calculating a first Z-fraction map corresponding to the cerebral blood flow image and a second Z-fraction map corresponding to the apparent diffusion coefficient image based on a predetermined brain image dataset of a healthy population; and processing the cerebral blood flow image, the first Z-fraction image, the apparent diffusion coefficient image and the second Z-fraction image by using a pre-trained classification model to obtain a regional segmentation image of the target object brain. According to the image area segmentation method, the ischemic penumbra segmentation is carried out through the combination of the cerebral blood flow image, the apparent diffusion coefficient image and the corresponding statistical characteristics, the information of an individual and the information compared with a healthy population are comprehensively considered, and the accuracy and the robustness of the ischemic penumbra segmentation can be improved.

Description

Image area segmentation method, device, equipment and storage medium

Technical Field

The present invention relates to the field of image processing technologies, and in particular, to an image region segmentation method, an image region segmentation apparatus, an image region segmentation device, and a storage medium.

Background

Ischemic stroke is a common cerebrovascular disease, seriously harms human health all the time, and has very high morbidity, disability rate and fatality rate. How to diagnose ischemic stroke quickly and effectively is an important research topic in current clinical work. The early detection of the ischemic stroke is very important, especially the correct detection of the hyperacute phase, can guide to take measures in time so as to reduce the brain cell necrosis of the ischemic penumbra area around the infarction focus, and has important effect on the diagnosis and treatment of the ischemic stroke.

In the prior art, brain parenchymal Imaging is commonly used for disease diagnosis, treatment effect evaluation and prognosis judgment, and among them, arterial Spin Labeling (ASL) perfusion Imaging technology and Diffusion Weighted Imaging (DWI) technology are commonly used. Wherein the ASL technique is based on a calculated Cerebral Blood Flow (CBF) image compared with contralateral values, less than 40% of the image is an ischemic region, and the DWI technique is based on a calculated Apparent Diffusion Coefficient (ADC) image, less than 620mm ² The area of infarct is/s.

However, since the existing ASL technology and DWI technology only consider the information of the individual, the normal ADC value of some positions of the individual is less than 620mm ² And/s, the normal CBF value of some positions is lower, so that the accuracy of the judgment result of the prior art is poor. Also, the ASL technique may miss the possibility of bilateral ischemia by comparing with the contralateral value, and the DWI technique may make the determination result unstable by absolute threshold division.

Disclosure of Invention

In view of the above problems in the prior art, an object of the present invention is to provide an image region segmentation method, device, apparatus and storage medium, which can improve the accuracy and robustness of segmentation of ischemic penumbra.

In order to solve the above problem, the present invention provides an image region segmentation method, including:

acquiring a cerebral blood flow image and an apparent diffusion coefficient image of a brain of a target object;

calculating a first Z-fraction map corresponding to the cerebral blood flow image and a second Z-fraction map corresponding to the apparent diffusion coefficient image based on a predetermined brain image dataset of a healthy population;

and processing the cerebral blood flow image, the first Z-fraction image, the apparent diffusion coefficient image and the second Z-fraction image by using a pre-trained classification model to obtain a regional segmentation image of the target object brain.

Further, the acquiring the cerebral blood flow image and the apparent diffusion coefficient image of the brain of the target subject includes:

acquiring arterial spin labeling perfusion image data of a brain of a target object;

determining the cerebral blood flow image from the arterial spin labeling perfusion data;

acquiring diffusion weighted image data of a target object brain;

determining the apparent diffusion coefficient image from the diffusion weighted image data.

Further, the acquiring the cerebral blood flow image and the apparent diffusion coefficient image of the brain of the target subject further includes:

acquiring a T1 structural image of the target subject brain;

and respectively registering the cerebral blood flow image and the apparent diffusion coefficient image to a standard cerebral space based on the T1 structure image to obtain the cerebral blood flow image and the apparent diffusion coefficient image of the standard cerebral space.

Further, the brain image data set of the healthy population comprises a cerebral blood flow image set of a standard brain space and an apparent diffusion coefficient image set of the standard brain space;

the calculating a first Z-score map corresponding to the cerebral blood flow image and a second Z-score map corresponding to the apparent diffusion coefficient image based on a predetermined brain image dataset of a healthy population comprises:

calculating the Z fraction of each pixel point in the cerebral blood flow image based on the cerebral blood flow image set of the standard cerebral space to obtain a first Z fraction image;

and calculating the Z score of each pixel point in the apparent diffusion coefficient image based on the apparent diffusion coefficient image set of the standard brain space to obtain a second Z score image.

Further, the processing the cerebral blood flow image, the first Z-score map, the apparent diffusion coefficient image, and the second Z-score map by using a pre-trained classification model to obtain a region segmentation image of the brain of the target subject includes:

inputting the cerebral blood flow image, the first Z-score image, the apparent diffusion coefficient image and the second Z-score image into the classification model to obtain a classification image of the brain of the target object, wherein each pixel in the classification image is an area class identifier;

and fusing the classified image and the artery spin labeling perfusion image data to obtain a regional segmentation image of the brain of the target object.

Further, the arterial spin labeling perfusion image data comprises a proton density weighted image;

the fusing the classified image with the artery spin labeling perfusion image data to obtain the regional segmentation image of the brain of the target object comprises:

converting the classified image into an arterial spin labeling perfusion data space to obtain a classified image of the arterial spin labeling perfusion data space;

and fusing the classified image of the arterial spin labeling perfusion data space and the proton density weighted image to obtain a regional segmentation image of the brain of the target object.

Further, the method further comprises:

calculating the volume of a penumbra area and the volume of an ischemic area according to the area segmentation image;

calculating the miscompare ratio according to the volume of the penumbra area and the volume of the ischemic area.

Another aspect of the present invention provides an image region segmentation apparatus, including:

the image acquisition module is used for acquiring a cerebral blood flow image and an apparent diffusion coefficient image of the brain of the target object;

a first calculating module, configured to calculate, based on a predetermined brain image data set of a healthy person, a first Z-score map corresponding to the cerebral blood flow image and a second Z-score map corresponding to the apparent diffusion coefficient image;

and the image processing module is used for processing the cerebral blood flow image, the first Z-score image, the apparent diffusion coefficient image and the second Z-score image by using a pre-trained classification model to obtain a regional segmentation image of the brain of the target object.

Another aspect of the present invention provides an electronic device, including a processor and a memory, where at least one instruction or at least one program is stored in the memory, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the image region segmentation method as described above.

Another aspect of the present invention provides a computer-readable storage medium, in which at least one instruction or at least one program is stored, and the at least one instruction or the at least one program is loaded and executed by a processor to implement the image region segmentation method as described above.

Due to the technical scheme, the invention has the following beneficial effects:

according to the image region segmentation method provided by the embodiment of the invention, the first Z-score map corresponding to the cerebral blood flow image of the target brain and the second Z-score map corresponding to the apparent diffusion coefficient image are calculated on the basis of the brain image dataset of the predetermined healthy population, and then machine learning classification is carried out on the basis of the cerebral blood flow image, the first Z-score map, the apparent diffusion coefficient image and the second Z-score map, so that the region segmentation image of the ischemic penumbra is obtained. The method for segmenting the ischemic penumbra by combining the cerebral blood flow image, the apparent diffusion coefficient image and the corresponding statistical characteristics comprehensively considers the information of an individual and the information compared with a healthy population, and can improve the accuracy and the robustness of segmenting the ischemic penumbra.

Drawings

In order to more clearly illustrate the technical solution of the present invention, the drawings used in the embodiment or the description of the prior art will be briefly described below. It is obvious that the drawings in the following description are only some embodiments of the invention, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.

FIG. 1 is a schematic diagram of an implementation environment provided by an embodiment of the invention;

FIG. 2 is a flowchart of an image region segmentation method according to an embodiment of the present invention;

FIG. 3 is a schematic illustration of registration to standard brain space provided by an embodiment of the present invention;

FIG. 4 is a diagram illustrating a region segmentation result according to an embodiment of the present invention;

FIG. 5 is a schematic structural diagram of an image region segmentation apparatus according to an embodiment of the present invention;

fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in other sequences than those illustrated or described herein. Moreover, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, apparatus, article, or device that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or device.

Referring to the specification, fig. 1 is a schematic diagram illustrating an implementation environment provided by an embodiment of the invention. As shown in fig. 1, the implementation environment may include at least one medical scanning device 110 and a computer device 120, where the computer device 120 and each medical scanning device 110 may be directly or indirectly connected through wired or wireless communication, and the embodiment of the present invention is not limited thereto.

The computer device 120 may acquire medical image data of the brain of the target subject scanned by the medical scanning device 110, and determine a region segmentation image of the brain of the target subject by using the image region segmentation method provided by the embodiment of the present invention, so as to be referred by a doctor and guide to take measures in time. The medical scanning device 110 may be but not limited to a magnetic resonance imaging device, and the like, the computer device 120 may be but not limited to various servers, personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices, the server may be an independent server or a server cluster or a distributed system composed of a plurality of servers, and may also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content Delivery Networks (CDNs), and big data and artificial intelligence platforms.

It should be noted that fig. 1 is only an example. It will be appreciated by those skilled in the art that although only 1 medical scanning device 110 is shown in FIG. 1, it is not intended to limit embodiments of the present invention and that more or fewer medical scanning devices 110 may be included than shown.

Referring to the specification, fig. 2 illustrates a flow of an image region segmentation method according to an embodiment of the present invention, which may be applied to the computer device 120 in fig. 1, and specifically as shown in fig. 2, the method may include the following steps:

s210: a cerebral blood flow image and an apparent diffusion coefficient image of the brain of the target subject are obtained.

In the embodiment of the invention, the ASL perfusion imaging technology can be utilized to collect the image of the brain of the target object to obtain the corresponding ASL perfusion image data, the ASL perfusion image data is further utilized to calculate to obtain the CBF image, the magnetic resonance DWI technology can also be utilized to collect the image of the brain of the target object to obtain the corresponding DWI image data, and the DWI image data is further utilized to calculate to obtain the ADC image. Wherein the target object may be a patient who may have a brain disease.

Specifically, the acquiring of the cerebral blood flow image and the apparent diffusion coefficient image of the brain of the target subject may include:

acquiring diffusion weighted image data of a target subject brain;

In an embodiment of the present invention, the ASL perfusion image data may include a proton density weighted image (M0 image), a control image (control image, i.e. non-label image), and a label image (label image), and the CBF image may be calculated by the control image and the label image. Specifically, for an image acquired by using a pulse-type labeled ASL sequence, the CBF calculation formula for each voxel is:

for images acquired using a continuous or pseudo-continuous labeled ASL sequence, the CBF calculation formula for each voxel is:

wherein λ is brain tissue/blood flow distribution coefficient, size is 0.9ml/g, and SI is _control And SI _label Time-averaged signal intensity, T, for control and label images, respectively _1,blood Is the longitudinal relaxation time of blood in seconds, α is the labeling efficiency, α =0.85 for continuous or pseudo-continuous labeling ASL sequences and α =0.98 for pulsed labeling ASL sequences. SI (Standard interface) _PD Is the signal intensity of the proton density weighted image, and τ is the mark duration. The PLD is the post-marker delay time, i.e. how long after the marker data starts to be acquired, TI is the inversion time, note that TI is the ASL sequence term for pulse-type markers, and PLD is the ASL sequence term for pseudo-continuous markers.

TI ₁ The labeled blood flow in the ASL sequence, which is pulsed, reaches the temporal extent of the scan field. In the formulae (1) and (2), the values of λ and α are known, SI _control 、SI _label 、SI _PD Needs to be derived from the acquired image, T _1,blood τ, PLD, TI and TI ₁ Are obtained from the ASL sequence information of the magnetic resonance system.

In an embodiment of the invention, the DWI image data may comprise at least two DWI images with different diffusion sensitivities (b 0, b 1), from which the ADC image may be calculated. The specific calculation formula is as follows:

wherein, SI _b0 Is the signal strength, SI, of a B0-valued DWI image _b1 Is the signal strength of the DWI image at b1 value, the b1 value is greater than the b0 value (the b0 value may be 0), the ADC image may reflect the diffusion of protons.

It should be noted that, the above-mentioned methods for calculating a CBF image by using ASL perfusion image data and calculating an ADC image by using DWI image data may be executed by a computer device implementing the method provided by the embodiment of the present invention, or may be executed by other devices, and the obtained CBF image and ADC image are sent to the computer device, which is not limited by the embodiment of the present invention.

It should be noted that, in some possible embodiments, images of the brain of the target subject may also be acquired by other imaging technologies in the prior art (for example, a nuclear magnetic perfusion imaging technology, a CT perfusion imaging technology, and the like), and a corresponding cerebral blood flow image is calculated, which is not limited in the embodiment of the present invention.

In one possible embodiment, after the CBF image and the ADC image are acquired, they may be registered to standard brain space to simplify subsequent processing. Specifically, the acquiring the cerebral blood flow image and the apparent diffusion coefficient image of the brain of the target subject may further include:

acquiring a T1 structural image of the brain of the target object;

and respectively registering the cerebral blood flow image and the apparent diffusion coefficient image to a standard brain space based on the T1 structural image to obtain the cerebral blood flow image and the apparent diffusion coefficient image of the standard brain space.

In the embodiment of the present invention, the target object brain may be scanned again to obtain the T1 structure image. Because the CBF image and the ADC image are parameter maps, which are different from the structural map, and if they are registered directly to the standard brain space, the error will be large, so the original images are generally used for registration (for example, using the M0 image in the ASL perfusion image data and the b0 image in the DWI image data), the conversion parameters from the ASL perfusion data space and the DWI data space to the standard brain space are obtained, and then the CBF image and the ADC image are converted to the standard brain space according to the obtained conversion parameters, so as to obtain the CBF image and the ADC image of the standard brain space. This is because the ADC image is calculated from DWI data (CBF image is calculated from ASL perfusion data) and therefore their space is consistent (i.e. pixels at the same location are identical), so that after the b0 image (or M0 image) is registered to the standard brain space, the ADC image (or CBF image) can be transformed to the standard brain space using the same transformation parameters.

Specifically, as shown in fig. 3, the b0 image in the DWI image data and the M0 image in the ASL perfusion image data may be registered to the T1 structure image, respectively, to obtain a first transformation parameter and a second transformation parameter for transforming to the T1 structure image, the T1 structure image may be registered to the MNI image in the standard brain space, to obtain a third transformation parameter for transforming the T1 structure image to the standard brain space, and finally, the transformation parameter for transforming from the DWI data space to the standard brain space may be determined based on the first transformation parameter and the third transformation parameter, and the transformation parameter for transforming from the ASL perfusion data space to the standard brain space may be determined based on the second transformation parameter and the third transformation parameter.

S220: calculating a first Z-score map corresponding to the cerebral blood flow image and a second Z-score map corresponding to the apparent diffusion coefficient image based on a predetermined brain image dataset of a healthy population.

In the embodiment of the invention, the statistical characteristics of each pixel point in the CBF image and the ADC image, namely the difference (i.e., Z score) between the CBF value of each pixel point in the CBF image and the average value of the healthy population, and the difference (i.e., Z score) between the ADC value of each pixel point in the ADC image and the average value of the healthy population, can be calculated based on the brain image dataset of the healthy population, and the target brain can be segmented by using the calculated statistical characteristic image (i.e., Z score).

In particular, the brain image dataset of the healthy population may comprise a cerebral blood flow image set of a standard brain space and an apparent diffusion coefficient image set of the standard brain space.

In practical application, ASL perfusion data and DWI data of a preset number of healthy people can be collected in advance, corresponding CBF images can be calculated for the ASL perfusion data of each healthy people, corresponding ADC images can be calculated for the DWI data of each healthy people, the calculated CBF images and the calculated ADC images are registered in a standard brain space, CBF images and ADC images of the standard brain space are obtained, and therefore a CBF image set and an ADC image set of the standard brain space of the healthy people can be finally established. The ASL perfusion data and the DWI data of the healthy population can be image data with high quality and without motion artifacts, so that CBF images and ADC images calculated based on the ASL perfusion data and the DWI data have no artifacts.

It should be noted that the preset number may be set according to actual needs, for example, the preset number may be set to at least 1000 examples, which is not limited in the embodiment of the present invention, and specific contents in step S210 may be referred to in the specific calculation method for the CBF image and the ADC image of the healthy people and the specific method for registering the CBF image and the ADC image in the standard brain space, which are not described herein again in the embodiment of the present invention.

In some possible embodiments, other imaging technologies (e.g., a nuclear magnetic perfusion imaging technology, a CT perfusion imaging technology, etc.) may also be used to acquire images of brains of a preset number of healthy people, and calculate to obtain corresponding CBF images, and then register the CBF images in the standard brain space to obtain CBF images in the standard brain space, so that a CBF image set in the standard brain space of the healthy people may be finally established.

It should be noted that, when the CBF image of the target brain and the CBF image of the brain of the healthy population are acquired and calculated by using the same imaging technology, the first Z-score map corresponding to the CBF image of the target brain may be calculated directly based on the CBF image set of the standard brain space of the healthy population. When different imaging technologies are adopted to acquire and calculate the CBF image of the target object brain and the CBF image of the healthy crowd brain, normalization processing may be performed on the CBF image of the target object brain and the CBF image of the healthy crowd CBF image set, and a first Z-score map corresponding to the CBF image of the target object brain is calculated based on the image data after the normalization processing.

Specifically, the calculating a first Z-score map corresponding to the cerebral blood flow image and a second Z-score map corresponding to the apparent diffusion coefficient image based on a predetermined brain image dataset of a healthy population may include:

In the embodiment of the invention, for each pixel point in the CBF image of the standard brain space, the difference (namely Z fraction) between the CBF value of the pixel point and the average value of healthy people can be calculated by adopting the following formula:

wherein CBF is the value of the current pixel point of the CBF image, F _i The value of a pixel point corresponding to the ith CBF image in the cerebral blood flow image set of the standard cerebral space of the healthy crowd is n, and the number of the CBF images in the cerebral blood flow image set of the standard cerebral space of the healthy crowd is n. Thus, a first Z score map corresponding to the CBF image of the target brain is generated, and the value of each pixel point in the map represents the Z score of the corresponding pixel point in the CBF image of the target brain in healthy people.

In the embodiment of the present invention, for each pixel point in the ADC image of the standard brain space, the difference (i.e., Z fraction) between the ADC value of the pixel point and the average value of the healthy population can be calculated by using the following formula:

where ADC is the value of the current pixel point of the ADC image, F _i The value of a pixel point corresponding to the ith ADC image in the standard brain space apparent diffusion coefficient image set of healthy people is obtained, and n is the number of the ADC images in the standard brain space apparent diffusion coefficient image set of healthy people. Thus, a second Z-fraction graph corresponding to the ADC image of the target object brain is generated, and the value of each pixel point in the graph represents the Z fraction of the corresponding pixel point in the ADC image of the target object brain in healthy people.

S230: and processing the cerebral blood flow image, the first Z score image, the apparent diffusion coefficient image and the second Z score image by using a pre-trained classification model to obtain a regional segmentation image of the brain of the target object.

In an embodiment of the present invention, a machine learning algorithm may be used to automatically segment the ischemic penumbra by combining the cerebral blood flow image, the first Z-fraction map, the apparent diffusion coefficient image, and the second Z-fraction map. The segmented image of the segmented region may include one or more of an infarct region, a semi-dark band region and a normal region, and the infarct region and the semi-dark band region together constitute an ischemic region.

Specifically, the processing the cerebral blood flow image, the first Z-fraction map, the apparent diffusion coefficient image, and the second Z-fraction map by using a pre-trained classification model to obtain a region segmentation image of the target brain may include:

In the embodiment of the present invention, the classification model may analyze and process the input cerebral blood flow image, the first Z-score map, the apparent diffusion coefficient image, and the second Z-score map, determine the region class to which each pixel of the target brain image belongs, and distinguish the region classes by different region class identifiers. The classification model may be obtained by training a preset neural network model through training image data labeled with a region class in advance, where the preset neural network model may include, but is not limited to, a classification model commonly used in the prior art, and may be, for example, a Random Forest (RF) model or a Support Vector Machine (SVM) model, and the like. The area categories may include an peduncle area, a semi-dark band area and a normal area, and different identifiers may be used to distinguish different areas, for example, 1 may be used to identify the peduncle area, 2 may be used to identify the semi-dark band area, 0 may be used to identify the normal area, and so on. It should be noted that the above distinguishing method is not limited to the embodiment of the present invention, and other identification methods may be used to distinguish different area types.

In the embodiment of the invention, in combination with reference to fig. 4 of the specification, in order to more intuitively display each region, the classified image and the image of the ASL perfusion data space may be fused to obtain a region segmentation image, and different regions are marked in different ways, so that a doctor can quickly determine information such as the position and range of each region. Illustratively, as shown in fig. 4, the region (a) is an infarct region, (b) is a semi-dark band region, and the entire region composed of (a) and (b) is an ischemic region, and the other regions are normal regions.

In one possible embodiment, the arterial spin label perfusion image data may comprise a proton density weighted image; the fusing the classified image with the arterial spin labeling perfusion image data to obtain a region segmentation image of the brain of the target subject may include:

In practical applications, the classified image may be fused with the original image (e.g., proton density weighted image) to obtain the region segmentation image. Specifically, a corresponding Z-score image is calculated by using a cerebral blood flow image and an apparent diffusion coefficient image of a standard brain space, the cerebral blood flow image of the standard brain space and a first Z-score image corresponding thereto, the apparent diffusion coefficient image of the standard brain space and a second Z-score image corresponding thereto are input into the classification model, and a classification image of the standard brain space can be obtained. The classified images can be converted into the ASL perfusion data space by performing an inverse operation using a conversion parameter from the ASL perfusion data space to the standard brain space, which is determined when the CBF images are registered to the standard brain space.

In practical applications, the classified image of the ASL perfusion data space may be superimposed on the proton density weighted image according to color labels, for example, the infarct area may be labeled with red, the semi-dark band area may be labeled with green, and different areas may be labeled with different colors, so that different areas may be visually displayed, and a doctor may quickly and accurately determine information such as a position and a range of each area.

In one possible embodiment, the method may further comprise the step of training the classification model using training image data of pre-labeled region classes.

Specifically, a training image data set may be obtained in advance, each training image in the training image data set may be labeled pixel by pixel, the classification regions are a normal region, an infarct region, and a semi-dark band region, and then a preset neural network model is trained by using the labeled training image data set to obtain the classification model. The preset neural network model may include, but is not limited to, a classification model commonly used in the prior art, for example, an RF model or an SVM, and the embodiment of the present invention is not limited thereto.

Specifically, in the process of training the classification model, a verification image data set may be further obtained, where the verification image data set is used to evaluate the performance of the classification model, that is, to test the performance of the trained model, and when the trained classification model meets the preset condition, the training is completed to obtain the classification model that can be used for classification.

In practical application, the corresponding CBF image, the ADC image, the Z-score map corresponding to the CBF image, and the Z-score map corresponding to the ADC image may be determined according to the verification image dataset, the trained classification model may be verified by using the four feature images, and when the trained classification model satisfies a preset condition, the classification model that may be used for classification may be obtained. The preset condition may be preset, for example, the correlation between the DICE coefficient or the volume value between the ischemia area/penumbra area obtained by the segmentation and the actual ischemia area/penumbra area may be set to be greater than a certain preset value, and the preset value may be set according to an actual situation, which is not limited in the embodiment of the present invention.

In one possible embodiment, the method may further comprise the steps of:

In the embodiment of the invention, after the region segmentation image is obtained, the number of pixel points of each region can be determined, and then the volume of each pixel point is multiplied to obtain the volume of each region. Since the ischemic region includes an infarct region and a penumbra region, the volume of the ischemic region is the sum of the volume of the infarct region and the volume of the penumbra region, and the miscompare ratio can be calculated by the following formula:

wherein, V ₁ Volume of the semi-dark zone, V ₂ Is the volume of the ischemic zone. By calculating the volume of the penumbra area, the volume of the ischemic area and the miscomparement ratio, the ischemic penumbra of the ischemic stroke of the target user can be effectively evaluated, and the method has guiding significance for diagnosis and treatment of the ischemic stroke.

In summary, according to the image region segmentation method in the embodiment of the present invention, a first Z-score map corresponding to a cerebral blood flow image of a target brain and a second Z-score map corresponding to an apparent diffusion coefficient image are calculated based on a predetermined brain image dataset of a healthy population, and machine learning classification is performed based on the cerebral blood flow image, the first Z-score map, the apparent diffusion coefficient image, and the second Z-score map, so as to obtain a region segmentation image of an ischemic penumbra. The method for segmenting the ischemic penumbra by combining the cerebral blood flow image, the apparent diffusion coefficient image and the corresponding statistical characteristics comprehensively considers the information of an individual and the information compared with a healthy population, and can improve the accuracy and the robustness of segmenting the ischemic penumbra.

Referring to the specification and fig. 5, a structure of an image region segmentation apparatus 500 according to an embodiment of the present invention is shown. As shown in fig. 5, the apparatus 500 may include:

an image acquisition module 510 for acquiring a cerebral blood flow image and an apparent diffusion coefficient image of a brain of a target subject;

a first calculating module 520, configured to calculate a first Z-score map corresponding to the cerebral blood flow image and a second Z-score map corresponding to the apparent diffusion coefficient image based on a predetermined brain image dataset of a healthy population;

an image processing module 530, configured to process the cerebral blood flow image, the first Z-score map, the apparent diffusion coefficient image, and the second Z-score map by using a pre-trained classification model, so as to obtain a region segmentation image of the brain of the target object.

In one possible embodiment, the apparatus 500 may further include a second calculating module for calculating a penumbra region volume and an ischemic region volume from the region segmentation image; calculating the miscompare ratio according to the volume of the penumbra area and the volume of the ischemic area.

It should be noted that, when the apparatus provided in the foregoing embodiment implements the functions thereof, the division of each functional module is merely used as an example, and in practical applications, the above function distribution may be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules, so as to complete all or part of the functions described above. In addition, the apparatuses provided in the above embodiments and the corresponding method embodiments belong to the same concept, and specific implementation processes thereof are detailed in the corresponding method embodiments, and are not described herein again.

An embodiment of the present invention further provides an electronic device, which includes a processor and a memory, where the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the image region segmentation method provided in the above method embodiment.

The memory may be used to store software programs and modules, and the processor may execute various functional applications and data processing by operating the software programs and modules stored in the memory. The memory can mainly comprise a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs needed by functions and the like; the storage data area may store data created according to use of the apparatus, and the like. Further, the memory may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device. Accordingly, the memory may also include a memory controller to provide the processor access to the memory.

In a specific embodiment, fig. 6 is a schematic diagram illustrating a hardware structure of an electronic device for implementing the image region segmentation method provided by the embodiment of the present invention, where the electronic device may be a computer terminal, a mobile terminal, or another device, and the electronic device may also participate in forming or including the image region segmentation apparatus provided by the embodiment of the present invention. As shown in fig. 6, the electronic device 600 may include one or more computer-readable storage media of the memory 610, one or more processing cores of the processor 620, an input unit 630, a display unit 640, a Radio Frequency (RF) circuit 650, a wireless fidelity (WiFi) module 660, and a power supply 670. Those skilled in the art will appreciate that the electronic device configuration shown in fig. 6 does not constitute a limitation of electronic device 600, and may include more or fewer components than shown, or some components in combination, or a different arrangement of components. Wherein:

the memory 610 may be used to store software programs and modules, and the processor 620 may execute various functional applications and data processing by operating or executing the software programs and modules stored in the memory 610 and calling data stored in the memory 610. The memory 610 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to use of the electronic device, and the like. In addition, the memory 610 may include high speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other volatile solid state storage device. Accordingly, the memory 610 may also include a memory controller to provide the processor 620 with access to the memory 610.

The processor 620 is a control center of the electronic device 600, connects various parts of the whole electronic device using various interfaces and lines, performs various functions of the electronic device 600 and processes data by operating or executing software programs and/or modules stored in the memory 610 and calling data stored in the memory 610, thereby integrally monitoring the electronic device 600. The Processor 620 may be a central processing unit, or may be other general-purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, a discrete hardware component, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The input unit 630 may be used to receive input numeric or character information and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control. In particular, the input unit 630 may include a touch sensitive surface 631 as well as other input devices 632. In particular, the touch-sensitive surface 631 may include, but is not limited to, a touch pad or touch screen, and the other input devices 632 may include, but are not limited to, one or more of a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a trackball, a mouse, a joystick, and the like.

The display unit 640 may be used to display information input by or provided to a user and various graphical user interfaces of an electronic device, which may be made up of graphics, text, icons, video, and any combination thereof. The Display unit 640 may include a Display panel 641, and optionally, the Display panel 641 may be configured in the form of a Liquid Crystal Display (LCD), an Organic Light-Emitting Diode (OLED), or the like.

The RF circuit 650 may be used for receiving and transmitting signals during information transmission and reception or during a call, and in particular, for receiving downlink information of a base station and then processing the received downlink information by the one or more processors 620; in addition, data relating to uplink is transmitted to the base station. In general, RF circuit 650 includes, but is not limited to, an antenna, at least one Amplifier, a tuner, one or more oscillators, a Subscriber Identity Module (SIM) card, a transceiver, a coupler, a Low Noise Amplifier (LNA), a duplexer, and the like. In addition, RF circuit 650 may also communicate with networks and other devices via wireless communications. The wireless communication may use any communication standard or protocol, including but not limited to Global System for Mobile communication (GSM), general Packet Radio Service (GPRS), code Division Multiple Access (CDMA), wideband Code Division Multiple Access (WCDMA), long Term Evolution (LTE), email, short Messaging Service (SMS), etc.

WiFi belongs to a short-distance wireless transmission technology, and the electronic equipment 600 can help a user to receive and send emails, browse webpages, access streaming media and the like through the WiFi module 660, and provides wireless broadband internet access for the user. Although fig. 6 shows the WiFi module 660, it is understood that it does not belong to the essential constitution of the electronic device 600, and may be omitted entirely as needed within the scope not changing the essence of the invention.

The electronic device 600 further comprises a power supply 670 (e.g., a battery) for supplying power to various components, which may preferably be logically connected to the processor 620 via a power management system, so as to manage charging, discharging, and power consumption management functions via the power management system. The power supply 670 may also include one or more dc or ac power sources, recharging systems, power failure detection circuitry, power converters or inverters, power status indicators, or any other component.

It should be noted that, although not shown, the electronic device 600 may further include a bluetooth module, and the like, which is not described herein again.

An embodiment of the present invention further provides a computer-readable storage medium, which can be disposed in an electronic device to store at least one instruction or at least one program for implementing an image region segmentation method, where the at least one instruction or the at least one program is loaded and executed by the processor to implement the image region segmentation method provided by the foregoing method embodiment.

Optionally, in an embodiment of the present invention, the storage medium may include, but is not limited to: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.

An embodiment of the invention also provides a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to cause the computer device to execute the image region segmentation method provided in the various alternative embodiments described above.

It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And that specific embodiments have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.

The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, as for the apparatus embodiment, since it is substantially similar to the method embodiment, the description is relatively simple, and for the relevant points, reference may be made to the partial description of the method embodiment.

It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims

1. An image region segmentation method, comprising:

processing the cerebral blood flow image, the first Z-score image, the apparent diffusion coefficient image and the second Z-score image by using a pre-trained classification model to obtain a regional segmentation image of the brain of the target object,

wherein the brain image data sets of the healthy people comprise a cerebral blood flow image set of a standard brain space and an apparent diffusion coefficient image set of the standard brain space,

based on the cerebral blood flow image set of the standard cerebral space, calculating the Z fraction of each pixel point in the cerebral blood flow image by adopting the following formula to obtain a first Z fraction image:

wherein cbf is the value of the current pixel point of the cerebral blood flow image, F _i The value of a pixel point corresponding to the ith cerebral blood flow image in the cerebral blood flow image set of the standard cerebral space of the healthy people is shown, and n is the number of the cerebral blood flow images in the cerebral blood flow image set of the standard cerebral space of the healthy people;

calculating the Z fraction of each pixel point in the apparent diffusion coefficient image by adopting the following formula based on the apparent diffusion coefficient image set of the standard brain space to obtain a second Z fraction image:

where adc is the value of the current pixel point of the apparent diffusion coefficient image, F _i The ith in the standard brain space apparent diffusion coefficient image set for healthy peopleAnd (3) the value of the pixel point corresponding to the apparent diffusion coefficient image, wherein n is the number of the apparent diffusion coefficient images in the apparent diffusion coefficient image set of the standard brain space of the healthy crowd.

2. The method of claim 1, wherein obtaining the cerebral blood flow image and the apparent diffusion coefficient image of the brain of the target subject comprises:

acquiring diffusion weighted image data of a target object brain;

3. The method of claim 2, wherein obtaining a cerebral blood flow image and an apparent diffusion coefficient image of the brain of the target subject further comprises:

acquiring a T1 structural image of the brain of the target object;

4. The method of claim 1, wherein the processing the cerebral blood flow image, the first Z-score map, the apparent diffusion coefficient image, and the second Z-score map using a pre-trained classification model to obtain the region segmentation image of the target subject brain comprises:

inputting the cerebral blood flow image, the first Z-fraction map, the apparent diffusion coefficient image and the second Z-fraction map into the classification model to obtain a classification image of the brain of the target object, wherein each pixel in the classification image is an area class identifier;

5. The method of claim 4, wherein the arterial spin labeling perfusion image data comprises a proton density weighted image;

the fusing the classified image with the arterial spin labeling perfusion image data to obtain a region segmentation image of the brain of the target object comprises:

and fusing the classified image of the artery spin labeling perfusion data space with the proton density weighted image to obtain a regional segmentation image of the brain of the target object.

6. The method of claim 1, further comprising:

7. An image region segmentation apparatus, comprising:

a first calculating module, configured to calculate a first Z-score map corresponding to the cerebral blood flow image and a second Z-score map corresponding to the apparent diffusion coefficient image based on a predetermined cerebral blood flow image data set of healthy people, where the cerebral blood flow image data set of healthy people includes a cerebral blood flow image set of a standard cerebral space and an apparent diffusion coefficient image set of a standard cerebral space,

calculating the Z fraction of each pixel point in the cerebral blood flow image by adopting the following formula based on the cerebral blood flow image set of the standard cerebral space to obtain the first Z fraction image:

/>

where adc is the value of the current pixel point of the apparent diffusion coefficient image, F _i The value of a pixel point corresponding to the ith apparent diffusion coefficient image in the standard brain space apparent diffusion coefficient image set of the healthy people is shown, and n is the number of the apparent diffusion coefficient images in the standard brain space apparent diffusion coefficient image set of the healthy people;

and the image processing module is used for processing the cerebral blood flow image, the first Z-fraction image, the apparent diffusion coefficient image and the second Z-fraction image by using a pre-trained classification model to obtain a regional segmentation image of the target object brain.

8. An electronic device, comprising a processor and a memory, wherein at least one instruction or at least one program is stored in the memory, and the at least one instruction or the at least one program is loaded by the processor and executed to implement the image region segmentation method according to any one of claims 1 to 6.

9. A computer-readable storage medium, wherein at least one instruction or at least one program is stored in the computer-readable storage medium, and the at least one instruction or the at least one program is loaded and executed by a processor to implement the image region segmentation method according to any one of claims 1 to 6.