CN114066827A

CN114066827A - Method and device for determining nuclear group area, computer equipment and readable storage medium

Info

Publication number: CN114066827A
Application number: CN202111289012.7A
Authority: CN
Inventors: 王晓雪; 葛传斌; 张旭; 周腾鹤
Original assignee: Wuhan United Imaging Zhirong Medical Technology Co Ltd
Current assignee: Wuhan United Imaging Zhirong Medical Technology Co Ltd
Priority date: 2021-11-02
Filing date: 2021-11-02
Publication date: 2022-02-18

Abstract

The application relates to a nuclear group region determination method, a nuclear group region determination device, a nuclear group region determination computer device and a readable storage medium, wherein the nuclear group region determination method comprises the steps of inputting a brain medical image and a brain atlas of a subject to be inspected into a deformation field determination model to obtain a registration deformation field; and carrying out deformation processing on the brain atlas according to the registration deformation field to obtain a deformation atlas, and determining a nuclear cluster area in the brain medical image according to the deformation atlas. The method for determining the nuclear cluster region does not need to depend on the experience of a doctor, and can improve the accuracy of determining the nuclear cluster region.

Description

Method and device for determining nuclear group area, computer equipment and readable storage medium

Technical Field

The present application relates to the field of image processing technologies, and in particular, to a method and an apparatus for determining a nuclear group region, a computer device, and a readable storage medium.

Background

For patients with neurological diseases such as Parkinson's disease, dystonia, depression and epilepsy, drug resistance can be generated by long-term drug treatment, and the patients can not walk down and move with limbs after stopping taking the drug, so that the life quality of the patients is seriously affected. Therefore, in clinical practice, Deep Brain Stimulation (DBS) is generally used for therapy, and DBS is a small pacemaker-like device placed under the skin of the chest to transmit electric signals to the Brain region for controlling movement through a very thin wire to help control the movement of the patient. Related researches show that the nuclear mass of the human brain is closely related to related neurological diseases such as Parkinson's disease and the like, and the pathological changes of the nuclear mass can cause various motor and cognitive disorders. In treatment with DBS, electrical signals are transmitted to the nucleus pulposus region through wires to control patient motion, and therefore, the nucleus pulposus region in a medical image of the brain needs to be located.

In the conventional technology, the nuclear group region is usually determined in the brain medical image depending on the experience of a doctor, which results in low accuracy of determining the nuclear group region.

Disclosure of Invention

In view of the foregoing, it is desirable to provide a nuclear group region determining method, apparatus, computer device and readable storage medium.

In a first aspect, an embodiment of the present application provides a method for determining a nuclear cluster region, including:

inputting a brain medical image and a brain atlas of an object to be checked into a deformation field determination model to obtain a registration deformation field; the brain atlas comprises a probability map of a nuclear cluster region and a brain medical image template, and the deformation field determination model is obtained by training a deep learning network according to a brain atlas set and a brain medical image sample;

carrying out deformation processing on the brain atlas according to the registration deformation field to obtain a deformation atlas;

and determining a nuclear cluster region in the brain medical image according to the deformation map.

In one embodiment, before inputting the medical brain image and the brain atlas of the object to be examined into the deformation field determination model and obtaining the registered deformation field, the method for determining the nuclear group region further comprises the following steps:

acquiring an original brain medical image of a subject to be examined;

inputting an original brain medical image into an image preprocessing model to obtain a skull separation mask image and a bias field image;

and determining a brain medical image according to the skull separation mask image and the offset field image.

In one embodiment, an image pre-processing model includes an encoder, a first decoder, and a second decoder; inputting an original brain medical image into an image preprocessing model to obtain a skull separation mask image and a bias field image, wherein the method comprises the following steps:

inputting an original brain medical image into the encoder to obtain L-layer characteristics, wherein L is an integer greater than or equal to zero;

inputting the L-layer characteristics into a first decoder to obtain a skull separation mask image;

and inputting the L-layer characteristics into a second decoder to obtain the offset field image.

acquiring a brain atlas set; wherein, the brain atlas set comprises brain atlases corresponding to examination objects of different age groups;

and determining the brain atlas corresponding to the object to be checked in the brain atlas set according to the age of the object to be checked.

In one embodiment, obtaining a set of brain atlases includes:

acquiring a plurality of brain medical image sample sets, wherein each brain medical image sample set comprises a plurality of brain medical image samples corresponding to examination objects of the same age group, and each brain medical image sample comprises a labeled nucleus area;

for each brain medical image sample set, determining a brain atlas corresponding to the brain medical image sample set according to a plurality of brain medical image samples in the brain medical image sample set;

and taking a plurality of brain atlases corresponding to the plurality of brain medical image sample sets as a brain atlas set.

In one embodiment, determining a brain atlas corresponding to the brain medical image sample set according to a plurality of brain medical image samples in the brain medical image sample set includes:

registering each brain medical image sample in the brain medical image sample set to a standard space based on an affine transformation method;

determining a probability map of the nuclear group region according to the distribution of the nuclear group region in the standard space in each brain medical image sample;

and determining a brain atlas corresponding to the brain medical image sample set according to the acquired brain medical image template and the probability map of the nuclear group region to obtain a brain atlas set.

In one embodiment, the method for determining the nuclear group region further includes:

and correcting the nuclear group region by adopting a region growing method or a boundary correcting method.

In a second aspect, an embodiment of the present application provides a nuclear cluster region determining apparatus, including:

a registered deformation field determining module used for inputting the brain medical image and the brain atlas of the object to be checked into the deformation field determining model to obtain a registered deformation field; the brain atlas comprises a probability map of the nuclear group region and a brain medical image template, and the deformation field determination model is obtained by training a deep learning network according to the brain atlas and the brain medical image sample;

the deformation map determining module is used for performing deformation processing on the brain map according to the registration deformation field to obtain a deformation map;

and the nuclear group area determining module is used for determining the nuclear group area in the brain medical image according to the deformation map.

In a third aspect, an embodiment of the present application provides a computer device, which includes a memory and a processor, where the memory stores a computer program, and the processor implements the steps of the method provided by the foregoing embodiment when executing the computer program.

In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the steps of the method provided by the foregoing embodiment.

The embodiment of the application provides a nuclear group region determining method, a nuclear group region determining device, computer equipment and a readable storage medium, wherein the method comprises the steps of inputting a brain medical image and a brain atlas of an object to be checked into a deformation field determining model to obtain a registration deformation field; and carrying out deformation processing on the brain atlas according to the registration deformation field to obtain a deformation atlas, and determining a nuclear cluster region in the brain medical image according to the deformation atlas. According to the nuclear group region determining method provided by the embodiment of the application, the brain atlas is subjected to deformation processing through the registration deformation field determined by the pre-trained deformation field determining model, so that the deformation atlas is obtained, and the nuclear group region in the brain medical image is determined according to the deformation atlas. The nuclear group area determined in the way does not need to depend on the experience of a doctor, and the influence caused by subjective judgment of the doctor can be avoided, so that the accuracy of determining the nuclear group area can be improved.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments or the conventional technologies of the present application, the drawings used in the description of the embodiments or the conventional technologies will be briefly introduced below, it is obvious that the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.

Fig. 1 is a schematic flowchart illustrating steps of a method for determining a nuclear cluster region according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a deformation field determination model training process provided in one embodiment of the present application;

fig. 3 is a schematic flowchart illustrating steps of another method for determining a nuclear cluster region according to an embodiment of the present application;

fig. 4 is a schematic flowchart illustrating steps of another method for determining a nuclear cluster region according to an embodiment of the present application;

FIG. 5 is a flowchart illustrating steps of another method for determining a nuclear cluster region according to an embodiment of the present disclosure;

FIG. 6 is a schematic diagram of an image pre-processing model training process provided in one embodiment of the present application;

FIG. 7 is a flowchart illustrating steps of another method for determining a nuclear cluster region according to an embodiment of the present disclosure;

FIG. 8 is a flowchart illustrating steps of another method for determining a nuclear cluster region according to an embodiment of the present disclosure;

FIG. 9 is a flowchart illustrating steps of another method for determining a nuclear cluster region according to an embodiment of the present disclosure;

FIG. 10 is a flowchart illustrating another exemplary process for bolus region determination and application provided in accordance with an embodiment of the present application;

fig. 11 is a schematic structural diagram of a nuclear bolus region determining apparatus according to an embodiment of the present application;

fig. 12 is a schematic structural diagram of a computer device according to an embodiment of the present application.

Detailed Description

In order to make the aforementioned objects, features and advantages of the present application more comprehensible, embodiments accompanying the present application are described in detail below with reference to the accompanying drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. This application is capable of embodiments in many different forms than those described herein and that modifications may be made by one skilled in the art without departing from the spirit and scope of the application and it is therefore not intended to be limited to the specific embodiments disclosed below.

The numbering of the components as such, e.g., "first", "second", etc., is used herein only to distinguish the objects as described, and does not have any sequential or technical meaning.

The method for determining the nuclear group area can be realized through computer equipment. Computer devices include, but are not limited to, control chips, personal computers, laptops, smartphones, tablets, and portable wearable devices. The method provided by the application can be realized through JAVA software and can also be applied to other software.

The following describes the technical solutions of the present application and how to solve the technical problems with the technical solutions of the present application in detail with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.

Fig. 1 provides a method for determining a nuclear cluster region according to an embodiment of the present application. The embodiment of the application specifically describes a method for determining a nuclear group region by taking computer equipment as an execution subject, and the method comprises the following steps:

step 100, inputting a brain medical image and a brain atlas of an object to be checked into a deformation field determination model to obtain a registration deformation field; the brain atlas comprises a probability map of the nuclear group region and a brain medical image template, and the deformation field determination model is obtained by training the deep learning network according to the brain atlas set and the brain medical image sample.

In order to perform the method for determining the nuclear mass region, the computer device needs to acquire a medical image of the brain of the subject to be examined. The brain medical image includes an imaging region corresponding to a nuclear mass to be located, i.e., a nuclear mass region. The medical brain image corresponding to the object to be examined can be pre-stored in the memory of the computer device, and the computer device can directly obtain the medical brain image from the memory when needed. The embodiment does not limit the specific method for acquiring the brain medical image as long as the function thereof can be realized. In an alternative embodiment, the brain medical image may be a magnetic resonance medical image, and specifically may be a magnetic resonance T1 image and/or a magnetic resonance T2 image, and the present embodiment does not limit the type and number of the brain medical images as long as the function thereof can be achieved.

The brain atlas may be pre-stored in a memory of the computer device, and the kind of the brain atlas may be various. Different examination objects correspond to different brain maps, and the brain maps comprise probability maps of the nuclear group areas and brain medical image templates. The brain atlas in this embodiment refers to a brain atlas corresponding to an object to be examined. The brain medical image template refers to a brain medical image corresponding to a probability map of a nuclear group region, and one probability map of a nuclear group region corresponds to one brain medical image template. The probability map of the nuclear group region carries information related to probability values of the nuclear group regions determined by the regions in the corresponding brain medical image template, and the position of the nuclear group region in the brain medical image can be determined according to the probability map of the nuclear group region.

After obtaining the brain medical image, the computer device obtains the brain atlas corresponding to the object to be checked from the memory, and inputs the brain medical image and the brain atlas of the object to be checked into the pre-trained deformation field determination model, so as to obtain the registration deformation field. The registration deformation field is used for registering the brain atlas and the brain medical image. The deformation field determination model is obtained by training a deep learning neural network by computer equipment according to the brain atlas set and the brain medical image samples. The brain atlas set comprises a plurality of brain atlases, and the brain medical image sample can be a brain medical image of the examination object acquired in advance. The present embodiment does not limit the specific process of determining the model by the deformation field, as long as the function thereof can be realized.

Referring to fig. 2, in an alternative embodiment, the training process of the deformation field determination model is as follows: inputting the brain medical image sample and a brain atlas corresponding to the brain medical image sample in the brain atlas set into a deep learning network to obtain a trained registration deformation field; carrying out deformation processing on the brain atlas according to the trained registration deformation field, comparing the brain atlas after deformation processing with the brain medical image sample, and adjusting parameters of the deep learning network according to the comparison result; and obtaining a new brain medical image sample, returning to the step of inputting the brain medical image sample and a brain map corresponding to the brain medical image sample in the brain map set into a deep learning network to obtain a trained registration deformation field, and obtaining the trained deep learning network, namely a deformation field determination model, after the comparison result of the brain map and the brain medical image sample after deformation processing meets a preset condition (the similarity between the brain map and the brain medical image sample is greater than a certain preset threshold value or the difference between the brain map and the brain medical image sample is less than another preset threshold value).

And 110, carrying out deformation processing on the brain atlas according to the registration deformation field to obtain a deformation atlas.

And step 120, determining a nuclear group area in the brain medical image according to the deformation map.

After the computer equipment obtains the registration deformation field, the brain atlas corresponding to the object to be inspected is subjected to deformation processing according to the registration deformation field, so that the deformation atlas can be obtained. The brain atlas comprises a probability map of the nuclear group region and a brain medical image template, the deformation atlas comprises a probability map of the deformed nuclear group region and a brain medical image template after deformation, the brain medical image template after deformation is mapped to the brain medical image, and the nuclear group region in the brain medical image can be determined according to the probability map of the deformed nuclear group region corresponding to the brain medical image template after deformation. The embodiment does not limit the specific process of performing deformation processing on the brain atlas and the specific method for determining the nucleus area, as long as the function of the brain atlas can be realized. Various types of nuclei are contained in medical images of the brain, such as: subthalamic nucleus and the medial globus pallidus. In this embodiment, the type and the number of the finally determined nuclear group regions are not limited, and the user can select the nuclear group regions according to actual needs.

The nuclear cluster region determining method provided by the embodiment of the application obtains a registration deformation field by inputting a brain medical image of an object to be checked and a brain atlas matched with the object to be checked into a deformation field determining model; and carrying out deformation processing on the brain atlas according to the registration deformation field to obtain a deformation atlas, and determining a nuclear cluster region in the brain medical image according to the deformation atlas. According to the nuclear group region determining method provided by the embodiment of the application, the brain atlas is subjected to deformation processing through the registration deformation field determined by the pre-trained deformation field determining model, so that the deformation atlas is obtained, and the nuclear group region in the brain medical image can be determined according to the deformation atlas. The nuclear group area determined in the way does not need to depend on the experience of a doctor, and the influence caused by subjective judgment of the doctor can be avoided, so that the accuracy of determining the nuclear group area can be improved.

Referring to fig. 3, in an embodiment, a possible implementation method for acquiring a brain medical image of a subject to be examined before inputting the brain medical image and a brain atlas of the subject to be examined into a deformation field determination model and obtaining a registered deformation field is provided, the method includes the steps of:

step 300, acquiring a raw brain medical image of a subject to be examined.

The original brain medical image is an image obtained by directly scanning an object to be examined. The raw brain medical image may be pre-stored in a memory of a computer device, which acquires the raw brain medical image directly in the memory when acquiring the brain medical image of the subject to be examined.

And 310, performing skull separation processing and bias field removal processing on the original brain medical image to obtain the brain medical image.

After the computer device obtains the original brain medical image, it processes it to obtain the final brain medical image. Specifically, the computer device performs skull separation processing on the original brain medical image, so that a scalp imaging region and a skull imaging region in the original brain medical image can be removed, interference of structures such as the skull and the scalp on a subsequent process of determining a nuclear group region is avoided, and accuracy of determining the nuclear group region in the brain medical image can be improved. The bias field is an uneven field generated by the scanner itself and the deviation during scanning, so that the original brain medical image may have a large brightness difference even between the same tissues. The computer equipment carries out bias field removal processing on the original brain medical image, and can ensure that the brightness of the same tissue in the obtained brain medical image is consistent, so that the accuracy of determining the nuclear mass region in the brain medical image can be improved. The present embodiment does not limit the specific method for performing the skull separation processing and the bias field removal processing on the brain medical image, as long as the function thereof can be achieved.

Referring to fig. 4, in an embodiment, a possible implementation method for performing skull separation processing and bias field removal processing on an original brain medical image to obtain a brain medical image is provided, and the method includes the following steps:

step 400, inputting an original brain medical image into an image preprocessing model to obtain a skull separation mask image and a bias field image; the image preprocessing model is obtained by training a deep learning network according to an original brain medical image sample.

After obtaining the brain medical image, the computer equipment inputs the brain medical image into a pre-trained image preprocessing model, and obtains a skull isolation mask image and a bias field image corresponding to the original brain medical image through the image preprocessing model. The image preprocessing model is obtained by training a deep learning network by computer equipment according to original brain medical image samples. The original brain medical image sample may be the same as the original brain medical image, and the description of the original brain medical image sample may refer to the above detailed description of the original brain medical image, which is not described herein again. The embodiment does not limit the specific training method of the image preprocessing model as long as the function of the image preprocessing model can be realized.

And step 410, determining a brain medical image according to the skull separation mask image and the offset field image.

After the computer equipment obtains the skull separation mask image and the offset field image according to the image preprocessing model, the computer equipment can firstly separate the skull and scalp imaging areas in the original brain medical image according to the skull separation mask image, and then remove the offset field in the brain medical image after the skull separation according to the offset field image to obtain the final brain medical image; the computer equipment can also remove the bias field in the original brain medical image according to the bias field image, and then separate the skull in the brain medical image after removing the bias field according to the skull separation mask image to obtain the final brain medical image. The embodiment does not limit the method for determining the brain medical image according to the skull separation mask image and the offset field image, as long as the function of the method can be realized.

Specifically, the computer device can separate the skull in the original brain medical image according to preset logical operation and arithmetic operation, and remove the bias field to obtain the brain medical image.

In an alternative embodiment, the computer device performs supervised training on the deep learning network, that is, the computer device stores therein a standard skull separation mask image, i.e., a first gold standard image, and a standard offset field image, i.e., a second gold standard image. Inputting an original brain medical image sample into a deep learning network by computer equipment to obtain a skull separation mask image and a bias field image which are subjected to initial pretreatment, comparing the skull separation mask image which is subjected to the initial pretreatment with a first gold standard image, comparing the bias field image which is subjected to the initial pretreatment with a second gold standard image, and adjusting parameters in the deep learning network according to a comparison result; and obtaining a new original brain medical image sample, returning to execute the step of inputting the original brain medical image sample into the deep learning network to obtain the initial preprocessed skull separation mask image and the initial preprocessed offset field image, and obtaining a trained deep learning network, namely an image preprocessing model, after the comparison result of the preprocessed skull separation mask image and the first gold standard image and the comparison result of the preprocessed offset field image and the second gold standard image both meet preset conditions (the similarity degree is greater than a certain preset threshold value or the difference degree is less than another preset threshold value).

In this embodiment, the pre-trained image pre-processing model is directly used to process the original brain medical image, so as to obtain the skull isolation mask image and the offset field image. This may improve the efficiency of determining medical images of the brain and thus the nuclear bolus region.

Referring to fig. 5, in one embodiment, an image pre-processing model includes an encoder, a first decoder, and a second decoder; the encoder comprises two outputs, one output being connected to the first decoder and the other output being connected to the second decoder. Inputting an original brain medical image into an image preprocessing model to obtain a skull separation mask image and a bias field image, wherein the steps of the possible implementation mode comprise the following steps:

step 500, inputting the original brain medical image into an encoder to obtain L-layer features, wherein the L-layer features are used for representing the features of the original brain medical image under L resolutions, and L is an integer greater than or equal to zero;

after the computer equipment inputs the original brain medical image into an encoder in the image preprocessing model, L-layer features of the original brain medical image under L resolutions can be extracted through the encoder, wherein one resolution corresponds to one layer of features, and L is an integer greater than or equal to zero. In an optional embodiment, when L is zero, the encoder extracts features of the original brain medical image at full resolution, and the larger the value of L, the smaller the resolution corresponding to L-layer features extracted by the encoder. The number of resolutions may be set by the user, but the embodiment is not limited thereto.

And step 510, inputting the L-layer characteristics into a first decoder to obtain a skull separation mask image.

After obtaining the L-layer features, the computer equipment inputs the L-layer features into a first decoder, and the first decoder processes the L-layer features to obtain a skull separation mask image.

In an alternative embodiment, the first decoder uses a skip connection to retain more information and the corresponding loss function of the first decoder uses a dice loss function.

And step 520, inputting the L-layer characteristics into a second decoder to obtain an offset field image.

After obtaining the L-layer characteristics, the computer device inputs the L-layer characteristics into a second decoder, and the second decoder processes the L-layer characteristics to obtain the offset field image.

In an alternative embodiment, the second decoder may predict the plurality of bias fields by processing each of the L-layer features. The larger the resolution of the feature is, the finer the offset field image obtained by predicting the feature is. The second decoder also uses a hopping connection to retain more information.

In an alternative embodiment, the image pre-processing model is trained from raw brain medical image samples. As shown in fig. 6, assuming that 3 layers of features, namely L is 3, are obtained after an original brain medical image sample passes through an encoder, and the 3 layers of features are referred to as L0 layer features, L1 layer features, and L2 layer features, respectively. Wherein, when L0 is 0, L0 layer features represent features of the original brain medical image sample at full resolution, when L1 is 1, L1 layer features represent features of the original brain medical image sample at first resolution, and when L2 is 2, L2 layer features represent features of the original brain medical image at second resolution. Wherein the full resolution, the first resolution and the second resolution are sequentially reduced. The L0 layer feature, the L1 layer feature and the L2 layer feature can be processed through a first decoder to obtain a predicted skull separation mask image, the image is compared with a first gold standard image, and parameters of an image preprocessing model are adjusted until the comparison result between the image and the first gold standard image meets a preset condition. The L0 layer feature, the L1 layer feature and the L2 layer feature can be processed through a second decoder to obtain a prediction bias field image, the image is compared with a second golden standard image, and parameters of an image preprocessing model are adjusted until the comparison result between the image and the second golden standard image meets a preset condition. In particular, the predicted offset field image obtained by processing the characteristics of the L0 layer by the second decoder is more refined and accurate.

The loss functions employed by the second decoder include histogram loss, regularization term loss, and mean square error loss.

The histogram loss function is used for calculating the difference of the voxel gray value histogram between the brain medical image obtained by removing the predicted offset field image obtained by the second decoder and a third gold standard image, wherein the third gold standard image is the brain medical image obtained by removing the standard offset field image. The formula for the histogram loss function is: l is_HC＝(1-d(H_g,H_p) Wherein, L_HCExpressing the histogram loss function, d (Hg, Hp) is expressed by

And calculating, wherein Hg represents the voxel gray value histogram distribution of the second gold standard image, and Hp represents the voxel gray value histogram distribution of the brain medical image after the prediction bias field image is removed. Hg (M) represents the number of voxels whose second gold standard image voxel gray-scale value falls within gray-scale level M,

representing the average voxel number of the second golden standard image when each voxel is uniformly within each cell of the histogram, Hp (M) representing the voxel number of the brain medical image with the voxel gray value within the gray level M after the prediction bias field image is removed,

the average voxel number of the brain medical image without the prediction bias field image is shown under the condition that each voxel uniformly falls in each interval of the histogram.

The regular term loss function is mainly used for ensuring the continuous smoothness of the bias field, and the specific calculation formula is as follows:

wherein,

which represents the bias field, is shown,

a regularized term loss function representing the bias field, p represents the bias field voxels,

representing the gradient of the predicted bias field. The gradient of the bias field calculated according to the above formula may control the degree of smoothing of the bias field.

The mean square error loss function is calculated by

Wherein MSE (I)₁,I₂) Is represented by₁And I₂Mean square error loss function between, i.e. second golden standard image I₁And an offset field picture I outputted through the second decoder₂Mean square error function of between I_g(x, y) represents a pixel value of the second gold standard image at (x, y), I_m(x, y) denotes a pixel value of the offset field picture output by the second decoder at (x, y), and N denotes the number of pixels of the offset field picture.

In this embodiment, since the image pre-processing model needs to perform two tasks of skull separation and bias field removal, a dual-branch structure, i.e., a first decoder and a second decoder, is adopted. Because partial parameters adopted by the first decoder and the second decoder when processing L-layer characteristics can be shared, parameters can be reduced, the calculated amount is reduced, complementary information can be provided, and the finally obtained brain medical image is more accurate.

Referring to fig. 7, in one embodiment, before inputting a medical brain image and a brain atlas of a subject to be examined into a deformation field determination model to obtain a registered deformation field, the method for determining a nuclear mass region includes the steps of:

step 700, acquiring a brain atlas set; wherein, the brain atlas set comprises brain atlases corresponding to examination objects of different age groups.

And step 710, determining a brain atlas corresponding to the object to be checked in the brain atlas set according to the age of the object to be checked.

The brain atlas set comprises brain atlases corresponding to examination objects of different ages. The computer equipment judges the age of the object to be checked by acquiring the age of the object to be checked, and determines the brain atlas corresponding to the object to be checked in the age group as the brain atlas corresponding to the object to be checked.

In this embodiment, the brain atlas corresponding to the object to be examined is determined in the brain atlas set according to the age of the object to be examined, so that the problem that the finally determined nucleus area is inaccurate due to different brain structures of the objects to be examined in different ages can be avoided, that is, the accuracy of finally determining the nucleus area can be improved.

Referring to fig. 8, in an embodiment, a possible implementation manner of obtaining a brain atlas set is provided, which includes the steps of:

step 800, obtaining a plurality of brain medical image sample sets, wherein each brain medical image sample set comprises a plurality of brain medical image samples corresponding to examination objects of the same age group, and each brain medical image sample comprises a labeled nucleus region.

The computer device can obtain a plurality of brain medical image sample sets according to a plurality of brain medical image samples corresponding to the examination object of each age group. The brain medical image sample set comprises a plurality of brain medical image samples corresponding to examination objects of the same age group, each brain medical image sample comprises a labeled nucleus area, that is, the position of the nucleus area in the brain medical image sample is known. One examination correspondence may correspond to one brain medical image sample, or may correspond to a plurality of brain medical image samples, and there may be a plurality of examination subjects in the same age group. The computer device may finally acquire a plurality of brain medical image sample sets corresponding to examination subjects of different age groups. The brain medical image samples may be stored in a memory of a computer device, from which the computer device retrieves directly when needed. For the description of the brain medical image sample, reference may be made to the above detailed description of the brain medical image, which is not repeated herein.

Step 810, determining a brain atlas corresponding to the brain medical image sample set according to the plurality of brain medical image samples in the brain medical image sample set for each brain medical image sample set;

and step 820, taking a plurality of brain atlases corresponding to the plurality of brain medical image sample sets as a brain atlas set.

After obtaining the plurality of brain medical image sample sets, for each brain medical image sample set, the computer device may determine, according to the plurality of brain medical image samples in the brain medical image sample set, a brain atlas corresponding to the brain medical image sample set, and finally obtain a brain atlas corresponding to each brain medical image sample set, that is, a brain atlas set. The present embodiment does not limit the specific method for determining the brain atlas corresponding to each brain medical image sample set, as long as the function of the brain medical image sample set can be achieved.

In one embodiment, a possible implementation manner of determining a brain atlas corresponding to a brain medical image sample set according to a plurality of brain medical image samples in the brain medical image sample set is proposed, and the steps shown in fig. 9 include:

and 900, registering each brain medical image sample in the brain medical image sample set to a standard space based on an affine transformation method.

After obtaining the plurality of brain medical image sample sets, the computer device registers the plurality of brain medical image samples in each brain medical image sample set to a standard space, such as a Montreal Neurological Institute (MNI) space, according to an affine transformation method. That is, pixel values of a plurality of brain medical image samples are mapped into a standard space. The present embodiment does not limit the specific registration method as long as the function thereof can be achieved.

And step 910, determining a probability map of the nuclear group region according to the distribution of the nuclear group region in the standard space in each brain medical image sample.

After the computer device registers the plurality of brain medical image samples in each brain medical image sample set into the standard space, by calculating the number of the nuclear group regions in each brain medical image sample distributed in the standard space, a probability map of the nuclear group region corresponding to each brain medical image sample set, that is, a probability map of the nuclear group region corresponding to the examination object of each age group can be determined.

And 920, determining a brain atlas corresponding to the brain medical image sample set according to the acquired brain medical image template and the probability map of the nuclear group region.

The computer device can determine a brain atlas corresponding to the brain medical image sample set, namely the brain atlas of the examination object under the age group corresponding to the brain medical image sample set according to the brain medical image template and the acquired probability map of the nuclear group region corresponding to each brain medical image sample set.

In the embodiment, the brain atlases of different ages are determined for the examination objects of different ages, so that more accurate nuclear group areas can be determined for the brain medical image samples of the examination objects of different ages.

In one embodiment, the method for determining the nuclear zone further comprises the steps of:

The computer equipment can realize the correction of the initial nucleus area by adopting any one of an area growing method and a boundary correction method. The region growing correction method is used for rapidly filling and removing the deviation region in the nucleus region by responding to the click operation of a user. The deviation region in the nuclear mass region is a region that is significantly different from the exact nuclear mass region. The boundary correction method is to represent the outline of the nuclear group area in a boundary point form, and correct the outline of the nuclear group area in response to the operation of dragging the boundary point by a user so as to obtain a corrected more accurate nuclear group area. When the boundary correction method is carried out, the boundary points near the dragging point can be adjusted in a self-adaptive mode, and a user does not need to check all the boundary points of the nuclear group area, so that the workload of the user can be reduced, and the efficiency of determining and correcting can be improved.

In an alternative embodiment, after determining the nuclear mass region, the computer device may further plan the electrode implantation path according to the nuclear mass region. Specifically, the central position of the nuclear mass region is used as an initial target point position, a voxel point of a skin region in a brain medical image is used as a needle inserting point set when an electrode is implanted, the safe distance from an implantation path formed from each needle inserting point to the initial target point to the brain tissue and the blood vessel region is calculated according to the blood vessel region and the brain tissue region in the brain medical image, and the optimal implantation path is determined by comprehensively considering the maximized safe distance and the minimized implantation path length and the maximized nuclear mass region covered by the electrode contact.

In another alternative implementation, the initial target and implantation path may be adjusted according to the application. Specifically, the initial target point may be extended inward by 5mm, the needle insertion point is used as a vertex, the space in the conical range within 3 ° of the implantation path at that time is used as an adjustment space of the implantation path, the minimum safe distance in the adjustment space is calculated, and the minimum safe distance is identified to play a role in prompting. And judging the planned implantation path by the user, if the judgment result is not satisfactory, readjusting the initial target point and the needle insertion point, and resetting a new minimum safe distance mark.

In an alternative embodiment, in practical application, a Computed Tomography (CT) image of an examination object is acquired before surgery, the CT image is registered with a brain medical image of a certain nuclear bolus region, and then pixels of the image are changed by a weighted average method of the pixels, so as to realize fusion of the CT image and the brain medical image and navigate subsequent surgery.

In one embodiment, an overall flow of the nuclear zone determination and application is provided as shown in fig. 10, and includes the steps of:

101, acquiring an original brain medical image of an object to be checked, specifically, performing magnetic resonance scanning on the head of the object to be checked before an operation, and scanning a T1 sequence and/or a T2 sequence to obtain a magnetic resonance image of the object to be checked, which is called an original brain medical image;

102, preprocessing an original brain medical image, specifically, inputting the original brain medical image into an image preprocessing model to obtain a skull separation mask image and a bias field image, and determining the brain medical image according to the skull separation mask image and the bias field image;

103, selecting a brain atlas corresponding to the object to be checked, specifically, determining the brain atlas corresponding to the object to be checked in the brain atlas set according to the age of the object to be checked;

104, determining a nuclear cluster region in the brain medical image according to the brain medical image and the brain atlas of the object to be checked, specifically, inputting the brain medical image and the brain atlas of the object to be checked into a deformation field determination model to obtain a registration deformation field, performing deformation processing on the brain atlas according to the registration deformation field to obtain a deformation atlas, and determining the nuclear cluster region in the brain medical image according to the deformation atlas;

105, correcting the nuclear group region by adopting a region growing method or a boundary correction method;

step 106, planning an electrode implantation path according to the nuclear cluster area;

step 107, acquiring a three-dimensional brain medical image with a mark point of an object to be inspected, specifically, performing computed tomography or magnetic resonance scanning on the object to be inspected in the morning to obtain a computed tomography image or a magnetic resonance image with a bone mark point, and taking the computed tomography image or the magnetic resonance image as the three-dimensional brain medical image with the mark point;

108, carrying out image fusion on the three-dimensional brain medical image with the mark points and the brain medical image with the nuclear group area determined, specifically, carrying out registration on the two images, then changing pixels of the images by a pixel weighted average method to realize image fusion, further determining an operation path and navigating for a subsequent operation;

step 109, the determined surgical path is saved as a surgical plan.

It should be understood that, although the steps in the flowcharts in the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in the figures may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performance of the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

Referring to fig. 11, an embodiment of the present application provides a nuclear mass region determining apparatus 10, which includes a registration deformation field determining module 11, a deformation atlas determining module 12, and a nuclear mass region determining module 13. Wherein,

the registration deformation field determining module 11 is configured to input the brain medical image and the brain atlas of the object to be examined into the deformation field determining model to obtain a registration deformation field; the brain atlas comprises a probability map of the nuclear group region and a brain medical image template, and the deformation field determination model is obtained by training a deep learning network according to the brain atlas and the brain medical image sample;

the deformation map determining module 12 is used for performing deformation processing on the brain map according to the registration deformation field to obtain a deformation map;

the nuclear group area determining module 13 is configured to determine a nuclear group area in the brain atlas medical image according to the deformation atlas.

In one embodiment, the nuclear mass region determining apparatus 10 further comprises a first acquisition unit for acquiring a raw brain medical image of the subject to be examined; the processing unit is used for carrying out skull separation processing and bias field removal processing on the original brain medical image to obtain the brain medical image.

In one embodiment, the processing unit is specifically configured to input the original brain medical image into the image preprocessing model, so as to obtain a skull separation mask image and a bias field image; the image preprocessing model is obtained by training a deep learning network according to an original brain medical image sample; and determining a brain medical image according to the skull separation mask image and the bias field image.

In an embodiment, the processing unit is further specifically configured to input the original brain medical image into the encoder, to obtain L-layer features, where the L-layer features are used to characterize features of the original brain medical image at L resolutions, and L is an integer greater than or equal to zero; inputting the L-layer characteristics into a first decoder to obtain a skull separation mask image; and inputting the L-layer characteristics into a second decoder to obtain the offset field image.

In one embodiment, the apparatus 10 further comprises a second obtaining unit for obtaining a set of brain atlases; wherein, the brain atlas set comprises brain atlases corresponding to examination objects of different age groups; and determining the brain atlas corresponding to the object to be checked in the brain atlas set according to the age of the object to be checked.

In one embodiment, the nuclear mass region determining apparatus 10 further comprises a determining unit for acquiring a plurality of brain medical image sample sets; each brain medical image sample set comprises a plurality of brain medical image samples corresponding to examination objects of the same age group, and each brain medical image sample comprises a labeled nucleus area; for each brain medical image sample set, determining a brain atlas corresponding to the brain medical image sample set according to a plurality of brain medical image samples in the brain medical image sample set; and taking a plurality of brain atlases corresponding to the plurality of brain medical image sample sets as a brain atlas set.

In an embodiment, the determining unit is specifically configured to register each brain medical image sample of the set of brain medical image samples to a standard space based on an affine transformation method; determining a probability map of the nuclear group region according to the distribution of the nuclear group region in the standard space in each brain medical image sample; and determining a brain atlas corresponding to the brain medical image sample set according to the acquired brain medical image template and the probability map of the nuclear group region to obtain a brain atlas set.

In one embodiment, the bolus region determining apparatus 10 further comprises a correction unit for correcting the bolus region using a region growing method or a boundary correction method.

For the specific definition of the above-mentioned nuclear group region determining apparatus 10, the above definition of the nuclear group region determining method can be referred to, and is not described herein again. The respective modules in the bolus region determining apparatus 10 may be implemented in whole or in part by software, hardware, and combinations thereof. The above devices, modules or units may be embedded in hardware or independent from a processor in a computer device, or may be stored in a memory in the computer device in software, so that the processor can call and execute operations corresponding to the above devices or modules.

Referring to fig. 12, in one embodiment, a computer device is provided, and the computer device may be a server, and the internal structure thereof may be as shown in fig. 12. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing the brain atlas set, the deformation field determination model and the like. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer device is executed by a processor to implement a method of nuclear mass area determination.

Those skilled in the art will appreciate that the architecture shown in fig. 12 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, there is provided a computer device comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the following steps when executing the computer program:

In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring an original brain medical image of a subject to be examined; and carrying out skull separation processing and bias field removal processing on the original brain medical image to obtain the brain medical image.

In one embodiment, the processor, when executing the computer program, further performs the steps of: inputting an original brain medical image into an image preprocessing model to obtain a skull separation mask image and a bias field image; the image preprocessing model is obtained by training a deep learning network according to an original brain medical image sample; and determining a brain medical image according to the skull separation mask image and the bias field image.

In one embodiment, the processor, when executing the computer program, further performs the steps of: inputting the original brain medical image into an encoder to obtain L-layer features, wherein the L-layer features are used for representing the features of the original brain medical image under L resolutions, and L is an integer greater than or equal to zero; inputting the L-layer characteristics into a first decoder to obtain a skull separation mask image; and inputting the L-layer characteristics into a second decoder to obtain the offset field image.

In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring a brain atlas set; wherein, the brain atlas set comprises brain atlases corresponding to examination objects of different age groups; and determining the brain atlas corresponding to the object to be checked in the brain atlas set according to the age of the object to be checked.

In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring a plurality of brain medical image sample sets, wherein each brain medical image sample set comprises a plurality of brain medical image samples corresponding to examination objects of the same age group, and each brain medical image sample comprises a labeled nucleus area; for each brain medical image sample set, determining a brain atlas corresponding to the brain medical image sample set according to a plurality of brain medical image samples in the brain medical image sample set; and taking a plurality of brain atlases corresponding to the plurality of brain medical image sample sets as a brain atlas set.

In one embodiment, the processor, when executing the computer program, further performs the steps of: registering each brain medical image sample in the brain medical image sample set to a standard space based on an affine transformation method; determining a probability map of the nuclear group region according to the distribution of the nuclear group region in the standard space in each brain medical image sample; and determining a brain atlas corresponding to the brain medical image sample set according to the acquired brain medical image template and the probability map of the nuclear group region to obtain a brain atlas set.

In one embodiment, the processor, when executing the computer program, further performs the steps of: and correcting the nuclear group region by adopting a region growing method or a boundary correcting method.

In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:

In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring an original brain medical image of a subject to be examined; and carrying out skull separation processing and bias field removal processing on the original brain medical image to obtain the brain medical image.

In one embodiment, the computer program when executed by the processor further performs the steps of: inputting an original brain medical image into an image preprocessing model to obtain a skull separation mask image and a bias field image; the image preprocessing model is obtained by training a deep learning network according to an original brain medical image sample; and determining a brain medical image according to the skull separation mask image and the bias field image.

In one embodiment, the computer program when executed by the processor further performs the steps of: inputting the original brain medical image into an encoder to obtain L-layer features, wherein the L-layer features are used for representing the features of the original brain medical image under L resolutions, and L is an integer greater than or equal to zero; inputting the L-layer characteristics into a first decoder to obtain a skull separation mask image; and inputting the L-layer characteristics into a second decoder to obtain the offset field image.

In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring a brain atlas set; wherein, the brain atlas set comprises brain atlases corresponding to examination objects of different age groups; and determining the brain atlas corresponding to the object to be checked in the brain atlas set according to the age of the object to be checked.

In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring a plurality of brain medical image sample sets, wherein each brain medical image sample set comprises a plurality of brain medical image samples corresponding to examination objects of the same age group, and each brain medical image sample comprises a labeled nucleus area; for each brain medical image sample set, determining a brain atlas corresponding to the brain medical image sample set according to a plurality of brain medical image samples in the brain medical image sample set; and taking a plurality of brain atlases corresponding to the plurality of brain medical image sample sets as a brain atlas set.

In one embodiment, the computer program when executed by the processor further performs the steps of: registering each brain medical image sample in the brain medical image sample set to a standard space based on an affine transformation method; determining a probability map of the nuclear group region according to the distribution of the nuclear group region in the standard space in each brain medical image sample; and determining a brain atlas corresponding to the brain medical image sample set according to the acquired brain medical image template and the probability map of the nuclear group region to obtain a brain atlas set.

In one embodiment, the computer program when executed by the processor further performs the steps of: and correcting the nuclear group region by adopting a region growing method or a boundary correcting method.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. A method for determining a nuclear cluster region, comprising:

inputting a brain medical image and a brain atlas of an object to be checked into a deformation field determination model to obtain a registration deformation field; the brain atlas comprises a probability map of a nuclear group region and a brain medical image template, and the deformation field determination model is obtained by training a deep learning network according to a brain atlas set and a brain medical image sample;

and determining a nuclear group area in the brain medical image according to the deformation map.

2. The method of determining the nuclear bolus region according to claim 1, wherein before the medical brain image and the brain atlas of the subject to be examined are input into the deformation field determination model to obtain the registered deformation field, the method further comprises:

acquiring a raw brain medical image of the object to be examined;

inputting the original brain medical image into an image preprocessing model to obtain a skull separation mask image and a bias field image;

and determining the brain medical image according to the skull separation mask image and the bias field image.

3. The method of claim 2, wherein the image pre-processing model comprises an encoder, a first decoder, and a second decoder; inputting the original brain medical image into an image preprocessing model to obtain a skull separation mask image and a bias field image, wherein the method comprises the following steps:

inputting the original brain medical image into the encoder to obtain L-layer features, wherein L is an integer greater than or equal to zero;

inputting the L-layer characteristics into the first decoder to obtain the skull separation mask image;

and inputting the L-layer characteristics into the second decoder to obtain the bias field image.

4. The method of determining the nuclear bolus region according to claim 1, wherein before the medical brain image and the brain atlas of the subject to be examined are input into the deformation field determination model to obtain the registered deformation field, the method further comprises:

acquiring the brain atlas set; wherein the brain atlas set comprises brain atlases corresponding to examination objects of different age groups;

5. The method of determining a nuclei region according to claim 4, wherein the obtaining the set of brain atlases includes:

acquiring a plurality of brain medical image sample sets; wherein each brain medical image sample set comprises a plurality of brain medical image samples corresponding to examination objects of the same age group, and each brain medical image sample comprises a labeled nucleus area;

and taking a plurality of brain atlases corresponding to the plurality of brain medical image sample sets as the brain atlas set.

6. The method for determining the nuclear cluster region according to claim 5, wherein the determining the brain atlas corresponding to the brain medical image sample set according to the plurality of brain medical image samples in the brain medical image sample set comprises:

registering each of the brain medical image samples in the set of brain medical image samples to a standard space based on an affine transformation method;

determining a probability map of the nuclei region according to the distribution of the nuclei region in the standard space in each brain medical image sample;

and determining a brain atlas corresponding to the brain medical image sample set according to the acquired brain medical image template and the probability map of the nuclear group region.

7. The method of determining a nuclear mass region of claim 1, further comprising:

and correcting the nuclear group region by adopting a region growing method or a boundary correction method.

8. A nuclear bolus region determining apparatus, comprising:

a registered deformation field determining module used for inputting the brain medical image and the brain atlas of the object to be checked into the deformation field determining model to obtain a registered deformation field; the brain atlas comprises a probability map of a nuclear group region and a brain medical image template, and the deformation field determination model is obtained by training a deep learning network according to the brain atlas and the brain medical image sample;

the deformation map determining module is used for carrying out deformation processing on the brain map according to the registration deformation field to obtain a deformation map;

9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 7.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.