Right ventricle segmentation, cardiopulmonary analysis method and device, electronic equipment and storage medium

Abstract

The present disclosure relates to a right ventricle segmentation method, a heart-lung analysis method and device, an electronic device and a storage medium, and relates to the technical field of right ventricle analysis. The right ventricle segmentation method comprises: acquiring a chest image to be segmented, a first preset segmentation model and a second preset segmentation model; using the first preset segmentation model to perform overall segmentation of a left ventricle and a right ventricle on the chest image to be segmented to obtain a first segmentation image; using the second preset segmentation model to perform left ventricle segmentation on the chest image to be segmented to obtain a second segmentation image; and determining a right ventricle image based on the first segmentation image and the second segmentation image. The embodiment of the present disclosure can realize accurate segmentation of the right ventricle in the chest image, and further realize heart-lung analysis.

Description

Technical Field

[0001] This disclosure relates to the field of right ventricular analysis technology, and in particular to a right ventricular segmentation, cardiopulmonary analysis method and apparatus, electronic device and storage medium. Background Technology

[0002] The right ventricle is directly connected to the pulmonary artery, so assessing the structure and function of the right ventricle is crucial in diagnosing some lung diseases, such as chronic obstructive pulmonary disease (COPD). In COPD patients, the development of pulmonary hypertension leads to right ventricular dilation, right ventricular systolic and diastolic dysfunction, so the morphology and function of the right ventricle can aid in the diagnosis of COPD.

[0003] Assessment of the right ventricle typically requires segmentation in medical imaging. However, right ventricular segmentation presents several challenges: the right ventricle exhibits signal intensities similar to myocardium; its shape is complex, changing linearly from base to apex; segmenting images cut at the apex is difficult; and there are variations in intraventricular morphology and signal intensity among patients, potentially due to pathological changes. Simply put, the right ventricle is an irregularly shaped object, and its thin walls can sometimes blend in with surrounding tissues. Therefore, current clinical practice of manually segmenting the right ventricle is time-consuming and sometimes lacks precision. Thus, dynamic segmentation of the right ventricular region is needed to address these practical problems in visceral medicine. Summary of the Invention

[0004] This disclosure presents a technical solution for right ventricular segmentation, cardiopulmonary analysis method and apparatus, electronic equipment and storage medium.

[0005] According to one aspect of this disclosure, a method for right ventricular segmentation is provided, comprising:

[0006] Obtain the chest image to be segmented, the first preset segmentation model, and the second preset segmentation model;

[0007] Using the first preset segmentation model, the chest image to be segmented is divided into the left and right ventricles to obtain the first segmented image;

[0008] Using the second preset segmentation model, the left ventricle of the chest image to be segmented is segmented to obtain a second segmented image;

[0009] Based on the first segmented image and the second segmented image, the right ventricle image is determined.

[0010] Preferably, before obtaining the first preset segmentation model and the second preset segmentation model, the first preset segmentation model and the second preset segmentation model are trained respectively, and the training method includes:

[0011] Acquire augmented chest images and corresponding unenhanced chest images, first-label images of the left and right ventricles, and a second-label image of the left ventricle for training;

[0012] The enhanced chest image used for training and its corresponding unenhanced chest image are registered to obtain the registered image corresponding to the unenhanced chest image.

[0013] The registered image is segmented into a cardiac region to obtain a cardiac region image;

[0014] Using the heart region image and the first label image, a first preset segmentation model to be trained is obtained.

[0015] Furthermore, the second preset segmentation model is trained using the heart region image and the second label image to obtain the second preset segmentation model;

[0016] or,

[0017] Before obtaining the first preset segmentation model and the second preset segmentation model, the first preset segmentation model and the second preset segmentation model are trained respectively. The training method includes:

[0018] Acquire augmented chest images and corresponding unenhanced chest images, first-label images of the left and right ventricles, and a second-label image of the left ventricle for training;

[0019] The enhanced chest image used for training and its corresponding unenhanced chest image are respectively segmented into the cardiac region to obtain a first segmented image and a second segmented image.

[0020] The first segmented image and the second segmented image are registered to obtain a heart-registered image;

[0021] Using the heart registration image and the first label image, the first preset segmentation model to be trained is trained to obtain the first preset segmentation model;

[0022] Furthermore, the second preset segmentation model is trained using the heart registration image and the second label image to obtain the second preset segmentation model.

[0023] Preferably, the method for training a first preset segmentation model using the heart region image and the first label image to obtain the first preset segmentation model includes:

[0024] The heart region image and the first label image are reconstructed in three dimensions to obtain a first three-dimensional heart image and a first three-dimensional label image.

[0025] The first three-dimensional heart image and the first three-dimensional label image are projected from multiple set angles to obtain multiple second three-dimensional heart projection images and corresponding multiple second three-dimensional label projection images.

[0026] Using the heart region image, the first label image, the plurality of second three-dimensional heart projection images and the corresponding plurality of second three-dimensional label projection images, the first preset segmentation model to be trained is trained to obtain the first preset segmentation model;

[0027] And / or,

[0028] The method for training a second preset segmentation model using the heart region image and the second label image to obtain the second preset segmentation model includes:

[0029] The heart region image and the second label image are reconstructed in three dimensions to obtain a third three-dimensional heart image and a third three-dimensional label image.

[0030] The third three-dimensional heart image and the third three-dimensional label image are projected according to multiple set angles to obtain multiple fourth three-dimensional heart projection images and corresponding multiple fourth three-dimensional label projection images.

[0031] Using the heart region image, the second label image, the plurality of fourth three-dimensional heart projection images and the corresponding plurality of fourth three-dimensional label projection images, the first preset segmentation model to be trained is trained to obtain the second preset segmentation model;

[0032] And / or,

[0033] The method for training a first preset segmentation model using the cardiac registration image and the first label image to obtain the first preset segmentation model includes:

[0034] The cardiac registration image and the first label image are reconstructed in three dimensions to obtain a first three-dimensional cardiac registration image and a first three-dimensional label image.

[0035] The first three-dimensional cardiac registration image and the first three-dimensional label image are projected from multiple set angles to obtain multiple second three-dimensional cardiac projection registration images and corresponding multiple second three-dimensional label projection images.

[0036] Using the cardiac registration image, the first label image, the plurality of second three-dimensional cardiac projection registration images and the corresponding plurality of second three-dimensional label projection images, the first preset segmentation model to be trained is trained to obtain the first preset segmentation model;

[0037] And / or,

[0038] The method for training a second preset segmentation model using the cardiac registration image and the second label image to obtain the second preset segmentation model includes:

[0039] The heart region image and the second label image are reconstructed in three dimensions to obtain a third three-dimensional heart registration image and a third three-dimensional label image.

[0040] The third three-dimensional cardiac registration image and the third three-dimensional label image are projected according to multiple set angles to obtain multiple fourth three-dimensional cardiac projection registration images and corresponding multiple fourth three-dimensional label projection images.

[0041] Using the heart region image, the second label image, the multiple fourth three-dimensional heart projection registration images and the corresponding multiple fourth three-dimensional label projection images, the first preset segmentation model to be trained is trained to obtain the second preset segmentation model.

[0042] Preferably, the method for determining the right ventricular image based on the first segmented image and the second segmented image includes: subtracting the second segmented image from the first segmented image to determine the right ventricular image; and / or, the chest image to be segmented is a non-enhanced chest image.

[0043] According to one aspect of this disclosure, a cardiopulmonary analysis method is provided, comprising: obtaining multiple right ventricular images using the aforementioned right ventricular segmentation method;

[0044] The gender and chronic obstructive pulmonary disease (COPD) grade corresponding to the multiple right ventricular images were determined respectively;

[0045] Based on the gender and the classification of chronic obstructive pulmonary disease, the images of the right ventricle were grouped.

[0046] Each group of right ventricular images was then identified for multiple morphological information.

[0047] Based on the aforementioned morphological information, the relationship between different genders and chronic obstructive pulmonary disease (COPD) grading was analyzed.

[0048] And / or,

[0049] Obtain lung function data and a preset classification model corresponding to the multiple right ventricular images;

[0050] Determine multiple morphological information corresponding to the multiple right ventricular images respectively;

[0051] Using the preset classification model, dyspnea is identified based on the multiple morphological information and the lung function data.

[0052] Preferably, the method for determining the chronic obstructive pulmonary disease (COPD) classification corresponding to the multiple right ventricular images includes:

[0053] Obtain chest images corresponding to the multiple right ventricular images and a preset grading model;

[0054] Using the preset grading model, the chest images corresponding to the multiple right ventricular images are graded to determine the chronic obstructive pulmonary disease grade corresponding to the multiple right ventricular images;

[0055] And / or,

[0056] The method for analyzing the relationship between different genders and chronic obstructive pulmonary disease (COPD) grades based on the aforementioned morphological information includes:

[0057] Get the set significance value;

[0058] Significance analysis was performed on the multiple morphological information corresponding to different genders and chronic obstructive pulmonary disease grades to obtain multiple significance values;

[0059] By selecting morphological information that is less than or equal to the set significance value from among the multiple significance values, morphological information that affects different genders and chronic obstructive pulmonary disease grading is obtained;

[0060] And / or, it also includes: acquiring the image features corresponding to the lung images in the chest images corresponding to the multiple right ventricular images;

[0061] Using the preset recognition model, dyspnea is identified based on the multiple morphological information, the lung function data, and the imaging features.

[0062] According to one aspect of this disclosure, a right ventricular segmentation device is provided, comprising:

[0063] The first acquisition unit is used to acquire the chest image to be segmented, the first preset segmentation model, and the second preset segmentation model;

[0064] The first segmentation unit is used to perform overall segmentation of the left and right ventricles of the chest image to be segmented using the first preset segmentation model to obtain the first segmented image.

[0065] The second segmentation unit is used to segment the left ventricle of the chest image to be segmented using the second preset segmentation model to obtain a second segmented image;

[0066] The first determining unit is used to determine the right ventricular image based on the first segmented image and the second segmented image.

[0067] According to one aspect of this disclosure, a cardiopulmonary analyzer is provided, comprising:

[0068] The second determining unit is used to determine the gender and chronic obstructive pulmonary disease grade corresponding to the multiple right ventricular images respectively;

[0069] A grouping unit is used to group the right ventricular images based on the gender and the grade of chronic obstructive pulmonary disease.

[0070] The third determining unit is used to determine multiple morphological information corresponding to the grouped multiple right ventricular images respectively;

[0071] The analysis unit is used to analyze the relationship between different genders and chronic obstructive pulmonary disease (COPD) grades based on the multiple morphological information.

[0072] And / or,

[0073] The second acquisition unit is used to acquire lung function data and a preset classification model corresponding to the multiple right ventricular images;

[0074] The fourth determining unit is used to determine multiple morphological information corresponding to the multiple right ventricular images respectively;

[0075] The identification unit is used to identify dyspnea based on the preset classification model, the multiple morphological information and the lung function data.

[0076] According to one aspect of this disclosure, an electronic device is provided, comprising:

[0077] processor;

[0078] Memory used to store processor-executable instructions;

[0079] The processor is configured to perform the above-described right ventricular segmentation and / or cardiopulmonary analysis methods.

[0080] According to one aspect of this disclosure, a computer-readable storage medium is provided that stores computer program instructions thereon, which, when executed by a processor, implement the above-described right ventricular segmentation and / or cardiopulmonary analysis methods.

[0081] In this embodiment, the overall anatomical structure of the left and right ventricles, as well as the anatomical structure of the right ventricle, are considered, thereby achieving accurate segmentation of the right ventricle in chest images and enabling cardiopulmonary analysis. This solves the current problem of difficulty in segmenting the right ventricle in chest images and addresses the difficulty in effectively combining cardiopulmonary information for further analysis of chronic obstructive pulmonary disease.

[0082] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure.

[0083] Other features and aspects of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0084] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the specification, serve to illustrate the technical solutions of this disclosure.

[0085] Figure 1 A flowchart of a right ventricular division method according to an embodiment of the present disclosure is shown;

[0086] Figure 2 A schematic diagram showing the right ventricular segmentation result according to an embodiment of the present disclosure is shown;

[0087] Figure 3 This is a block diagram illustrating an electronic device 800 according to an exemplary embodiment;

[0088] Figure 4 This is a block diagram illustrating an electronic device 1900 according to an exemplary embodiment. Detailed Implementation

[0089] Various exemplary embodiments, features, and aspects of this disclosure will now be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0090] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0091] In this document, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Furthermore, the term "at least one" in this document means any combination of at least two of any one or more elements. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C.

[0092] Furthermore, to better illustrate this disclosure, numerous specific details are set forth in the following detailed description. Those skilled in the art will understand that this disclosure can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art have not been described in detail in order to highlight the main points of this disclosure.

[0093] It is understood that the various method embodiments mentioned above in this disclosure can be combined with each other to form combined embodiments without violating the principle and logic. Due to space limitations, this disclosure will not elaborate further.

[0094] In addition, this disclosure also provides a right ventricular segmentation and / or cardiopulmonary analysis device, electronic device, computer-readable storage medium, and program, all of which can be used to implement any of the right ventricular segmentation and / or cardiopulmonary analysis methods provided in this disclosure. The corresponding technical solutions and descriptions are described in the corresponding section of the method and will not be repeated here.

[0095] Figure 1 A flowchart illustrating a right ventricular division method according to an embodiment of this disclosure is shown. Figure 1 As shown, the right ventricle segmentation method includes: Step S101: acquiring a chest image to be segmented, a first preset segmentation model, and a second preset segmentation model; Step S102: using the first preset segmentation model to segment the chest image to be segmented into the left and right ventricles, obtaining a first segmented image; Step S103: using the second preset segmentation model to segment the chest image to be segmented into the left ventricle, obtaining a second segmented image; Step S104: determining the right ventricle image based on the first and second segmented images. This method enables accurate segmentation of the right ventricle in chest images, thereby achieving cardiopulmonary analysis and solving the current problem of difficulty in segmenting the right ventricle in chest images. It also addresses the difficulty in effectively combining cardiopulmonary information for further analysis of chronic obstructive pulmonary disease.

[0096] Step S101: Obtain the chest image to be segmented, the first preset segmentation model, and the second preset segmentation model.

[0097] In embodiments of this disclosure and other possible embodiments, the chest image may be a chest CT image, PET image, ultrasound image, MR image, DR image, or CT-PET image, etc. Chest CT images, PET images, ultrasound images, MR images, DR images, CT-PET images can be obtained using imaging equipment such as CT, ultrasound, MR, DR, CT-PET, etc.

[0098] In the embodiments of this disclosure and other possible embodiments, the first preset segmentation model and the second preset segmentation model can be a conventional heart segmentation model, or a heart segmentation model based on deep learning, such as a global segmentation model of the left and right ventricles and a left ventricular segmentation model based on U-Net, or U-ResNet or its improvements.

[0099] Step S102: Using the first preset segmentation model, the chest image to be segmented is divided into the left and right ventricles to obtain the first segmented image.

[0100] In the embodiments of this disclosure and other possible embodiments, the first preset segmentation model is a trained segmentation model used to perform overall segmentation of the left and right ventricles of the chest image to be segmented, to obtain the first segmented image.

[0101] Step S103: Using the second preset segmentation model, the left ventricle of the chest image to be segmented is segmented to obtain a second segmented image.

[0102] In the embodiments of this disclosure and other possible embodiments, the second preset segmentation model is a trained segmentation model used to segment the left ventricle of the chest image to be segmented, thereby obtaining a second segmented image.

[0103] In embodiments and other possible embodiments of this disclosure, a method or apparatus for training a first preset segmentation model and a second preset segmentation model is proposed to address the problem that patients with contraindications or severe cardiopulmonary diseases cannot undergo enhanced chest image scanning, making cardiac segmentation difficult. For example, enhanced chest images (enhanced chest images) can display the structure of the heart by intravenously injecting contrast agents. However, for some patients with contraindications to contrast agents and those without enhanced chest images (non-enhanced chest images), it is necessary to segment the right ventricle from the non-enhanced chest image.

[0104] In embodiments of this disclosure, before obtaining the first preset segmentation model and the second preset segmentation model, the first preset segmentation model and the second preset segmentation model are trained respectively. The training method includes: acquiring enhanced chest images and corresponding non-enhanced chest images, first label images of the left and right ventricles and a second label image of the left ventricle for training; registering the enhanced chest images and the corresponding non-enhanced chest images for training to obtain a registration image corresponding to the non-enhanced chest images; performing cardiac region segmentation on the registration images to obtain cardiac region images; training the first preset segmentation model to be trained using the cardiac region images and the first label images to obtain the first preset segmentation model; and training the second preset segmentation model to be trained using the cardiac region images and the second label images to obtain the second preset segmentation model. The enhanced chest images and corresponding non-enhanced chest images used for training are chest images of the same patient or subject.

[0105] In embodiments of this disclosure and other possible embodiments, two experienced radiologists labeled the left and right ventricles of multiple patients on original enhanced chest images, obtaining first labeled images of the left and right ventricles and a second labeled image of the left ventricle. First, semi-automatic labeling of the enhanced chest images was performed using Mimics software in the sagittal plane or other possible angles, with the left ventricle labeled as 1, the right ventricle as a whole labeled as 2, and other tissues labeled as 0 after binarization. Based on the above, the labels corresponding to the left and right ventricles were configured as the first labeled image, and the label corresponding to the left ventricle was configured as the second labeled image.

[0106] In the embodiments of this disclosure and other possible embodiments, the registration method for the enhanced chest image used for training and the corresponding non-enhanced chest image can be the Elastix registration module in 3D Slicer, or it can be the SIFT registration method, the 3D SIFT registration method, or the SURF registration method, etc.

[0107] In the embodiments of this disclosure and other possible embodiments, the cardiac region image (a whole cardiac segmentation image, which may include: left ventricle (LV), left ventricular myocardium (Myo), left atrium (LA), right atrium (RA), right ventricle (RV), pulmonary artery (PA), ascending artery (AA)) obtained by performing cardiac region segmentation on the registered image can be an existing cardiac segmentation model, such as a cardiac segmentation model based on U-Net, or U-ResNet or an improved version thereof. For example, a cardiac region image can be obtained using an artificial intelligence-based CT image cardiac segmentation method and system as described in application number 202210321078.8.

[0108] In the embodiments of this disclosure and other possible embodiments, the first and second preset segmentation models to be trained can use the same network structure, for example, both including: a U-Net backbone network, which obtains multiple feature maps after each convolution operation; feature mapping normalization is performed on the multiple feature maps respectively; and an activation function is used to activate the normalized multiple feature maps. In the embodiments of this disclosure, feature mapping normalization is performed on the multiple feature maps respectively to accelerate the convergence of the first and second preset segmentation models and to maintain the independence between each feature map. Simultaneously, the Leaky ReLU activation function can be used to activate the normalized multiple feature maps.

[0109] In embodiments of this disclosure and other possible embodiments, the method for training the first and second preset segmentation models to be trained further includes: calculating the loss of the feature map corresponding to each decoding process during the decoding process of the first and second preset segmentation models to be trained; and obtaining the total loss during the training process based on the loss of the feature map corresponding to each decoding process.

[0110] In embodiments of this disclosure, before obtaining the first preset segmentation model and the second preset segmentation model, the first preset segmentation model and the second preset segmentation model are trained respectively. The training method includes: acquiring an enhanced chest image and a corresponding non-enhanced chest image, a first label image of the left ventricle and the right ventricle, and a second label image of the left ventricle for training; performing cardiac region segmentation on the enhanced chest image and the corresponding non-enhanced chest image for training respectively to obtain a first segmentation image and a second segmentation image; registering the first segmentation image and the second segmentation image to obtain a cardiac registration image; training the first preset segmentation model to be trained using the cardiac registration image and the first label image to obtain a first preset segmentation model; and training the second preset segmentation model to be trained using the cardiac registration image and the second label image to obtain a second preset segmentation model.

[0111] Similarly, in embodiments of this disclosure and other possible embodiments, two experienced radiologists labeled the left and right ventricles of multiple patients in the original enhanced chest images, obtaining first labeled images of the left and right ventricles and a second labeled image of the left ventricle. First, semi-automatic labeling of the enhanced chest images was performed using Mimics software in the sagittal plane or other possible angles, with the left ventricle labeled as 1, the right ventricle as a whole labeled as 2, and other tissues labeled as 0 after binarization. Based on the above, the labels corresponding to the left and right ventricles were configured as the first labeled image, and the label corresponding to the left ventricle was configured as the second labeled image.

[0112] Similarly, in the embodiments of this disclosure and other possible embodiments, the first segmented image and the second segmented image obtained by segmenting the enhanced chest image used for training and its corresponding non-enhanced chest image into a cardiac region are both whole cardiac segmented images, which may include: left ventricle (LV), left ventricular myocardium (Myo), left atrium (LA), right atrium (RA), right ventricle (RV), pulmonary artery (PA), and ascending artery (AA). The cardiac segmentation model can be an existing cardiac segmentation model, such as a cardiac segmentation model based on U-Net, or U-ResNet, or an improved version thereof. For example, a cardiac region image can be obtained using an artificial intelligence-based CT image cardiac segmentation method and system as described in application number 202210321078.8.

[0113] Similarly, in the embodiments of this disclosure and other possible embodiments, the registration of the first segmented image and the second segmented image can be performed using the Elastix registration module in 3D Slicer, or the SIFT registration method, the 3DSIFT registration method, or the SURF registration method, etc.

[0114] Similarly, in the embodiments of this disclosure and other possible embodiments, the first and second preset segmentation models to be trained can use the same network structure, for example, both including: a U-Net backbone network, which obtains multiple feature maps after each convolution operation; feature mapping normalization is performed on the multiple feature maps respectively; and activation is applied to the normalized multiple feature maps using an activation function. In the embodiments of this disclosure, feature mapping normalization is performed on the multiple feature maps respectively to accelerate the convergence of the first and second preset segmentation models and to maintain the independence between each feature map. Simultaneously, the Leaky ReLU activation function can be used to activate the normalized multiple feature maps. Furthermore, in the embodiments of this disclosure and other possible embodiments, the method for training the first and second preset segmentation models to be trained further includes: calculating the loss of the feature map corresponding to each decoding step during the decoding process of the first and second preset segmentation models to be trained; and obtaining the total loss during the training process based on the loss of the feature map corresponding to each decoding step.

[0115] In the embodiments and other possible embodiments of this disclosure, a segmentation model based on a cardiac region image and its multi-angle projection images is proposed. This model adds corresponding multi-angle projection images to the original cardiac region image, increasing the robustness and universality of the segmentation model. Consequently, the segmentation model can simultaneously perform overall segmentation of the left and right ventricles and / or left ventricular segmentation in both 3D and 2D chest images. Simply put, if the multi-angles are only horizontal and vertical angles, the projection images can be the horizontal and vertical projection images of the cardiac region image or a registered cardiac region image. Specifically, the horizontal projection image is obtained by summing the column values ​​of the cardiac region image or the registered cardiac region image; the vertical projection image is obtained by summing the row values ​​of the cardiac region image or the registered cardiac region image.

[0116] In an embodiment of this disclosure, the method of training a first preset segmentation model using the heart region image and the first label image to obtain the first preset segmentation model includes: performing three-dimensional reconstruction on the heart region image and the first label image respectively to obtain a first three-dimensional heart image and a first three-dimensional label image; projecting the first three-dimensional heart image and the first three-dimensional label image from multiple set angles to obtain multiple second three-dimensional heart projection images and corresponding multiple second three-dimensional label projection images; and training the first preset segmentation model using the heart region image, the first label image, the multiple second three-dimensional heart projection images, and the corresponding multiple second three-dimensional label projection images to obtain the first preset segmentation model.

[0117] In the embodiments of this disclosure and other possible embodiments, the heart region image and the first label image can be reconstructed in three dimensions based on the reconstruction algorithm of ITK-snap software or other three-dimensional reconstruction algorithms to obtain a first three-dimensional heart image and a first three-dimensional label image. Then, the first three-dimensional heart image and the first three-dimensional label image are projected at multiple preset angles to obtain multiple second three-dimensional heart projection images and corresponding multiple second three-dimensional label projection images.

[0118] In the embodiments of this disclosure and other possible embodiments, the plurality of set angles can be set angles corresponding to sagittal plane, coronal plane cross-sectional views, and any angle, such as any angle can be 30°, 45°, 60° views in various possible orientations.

[0119] In an embodiment of this disclosure, the method of training a second preset segmentation model to be trained using the heart region image and the second label image to obtain the second preset segmentation model includes: performing three-dimensional reconstruction on the heart region image and the second label image respectively to obtain a third three-dimensional heart image and a third three-dimensional label image; projecting the third three-dimensional heart image and the third three-dimensional label image at multiple set angles to obtain multiple fourth three-dimensional heart projection images and corresponding multiple fourth three-dimensional label projection images; and training a first preset segmentation model to be trained using the heart region image, the second label image, the multiple fourth three-dimensional heart projection images and the corresponding multiple fourth three-dimensional label projection images to obtain the second preset segmentation model.

[0120] In the embodiments of this disclosure and other possible embodiments, the heart region image and the second label image can be reconstructed in three dimensions based on the reconstruction algorithm of ITK-snap software or other three-dimensional reconstruction algorithms to obtain a third three-dimensional heart image and a third three-dimensional label image. Furthermore, the first three-dimensional heart image and the first three-dimensional label image are projected at multiple preset angles to obtain multiple second three-dimensional heart projection images and corresponding multiple second three-dimensional label projection images.

[0121] In the embodiments of this disclosure and other possible embodiments, the plurality of set angles can be set angles corresponding to sagittal plane, coronal plane cross-sectional views, and any angle, such as any angle can be 30°, 45°, 60° views in various possible orientations.

[0122] In embodiments of this disclosure, the method for training a first preset segmentation model using the cardiac registration image and the first label image to obtain the first preset segmentation model includes:

[0123] The heart registration image and the first label image are reconstructed in three dimensions to obtain a first three-dimensional heart registration image and a first three-dimensional label image. The first three-dimensional heart registration image and the first three-dimensional label image are projected at multiple set angles to obtain multiple second three-dimensional heart projection registration images and corresponding multiple second three-dimensional label projection images. The heart registration image, the first label image, the multiple second three-dimensional heart projection registration images and the corresponding multiple second three-dimensional label projection images are used to train a first preset segmentation model to obtain a first preset segmentation model.

[0124] In the embodiments of this disclosure and other possible embodiments, the registered heart image and the first label image can be reconstructed in three dimensions based on the reconstruction algorithm of ITK-snap software or other three-dimensional reconstruction algorithms to obtain a first three-dimensional registered heart image and a first three-dimensional label image. Then, the first three-dimensional heart image and the first three-dimensional label image are projected at multiple set angles to obtain multiple second three-dimensional heart projection images and corresponding multiple second three-dimensional label projection images.

[0125] In the embodiments of this disclosure and other possible embodiments, the plurality of set angles can be set angles corresponding to sagittal plane, coronal plane cross-sectional views, and any angle, such as any angle can be 30°, 45°, 60° views in various possible orientations.

[0126] In an embodiment of this disclosure, the method of training a second preset segmentation model to be trained using the cardiac registration image and the second label image to obtain a second preset segmentation model includes: performing three-dimensional reconstruction on the cardiac registration image and the second label image respectively to obtain a third three-dimensional cardiac registration image and a third three-dimensional label image; projecting the third three-dimensional cardiac registration image and the third three-dimensional label image according to multiple set angles to obtain multiple fourth three-dimensional cardiac projection registration images and corresponding multiple fourth three-dimensional label projection images; and training a first preset segmentation model to be trained using the cardiac region image, the second label image, the multiple fourth three-dimensional cardiac projection registration images and the corresponding multiple fourth three-dimensional label projection images to obtain a second preset segmentation model.

[0127] In the embodiments of this disclosure and other possible embodiments, the registered heart image and the second label image can be reconstructed in three dimensions based on the reconstruction algorithm of ITK-snap software or other three-dimensional reconstruction algorithms to obtain a third three-dimensional registered heart image and a third three-dimensional label image. Then, the first three-dimensional heart image and the first three-dimensional label image are projected at multiple preset angles to obtain multiple second three-dimensional heart projection images and corresponding multiple second three-dimensional label projection images.

[0128] In the embodiments of this disclosure and other possible embodiments, the plurality of set angles can be set angles corresponding to sagittal plane, coronal plane cross-sectional views, and any angle, such as any angle can be 30°, 45°, 60° views in various possible orientations.

[0129] Step S104: Determine the right ventricular image based on the first segmented image and the second segmented image.

[0130] In embodiments of this disclosure, the method for determining a right ventricular image based on the first segmented image and the second segmented image includes: subtracting the second segmented image from the first segmented image to determine the right ventricular image; and / or, the chest image to be segmented is a non-enhanced chest image.

[0131] Figure 2 A schematic diagram showing the right ventricular segmentation result according to an embodiment of the present disclosure is provided. Figure 2 As shown, Figure 2 (a) shows the overall segmentation results of the right and left ventricles. Figure 2 (b) shows the division results of the left ventricle. Figure 2 (c) shows the segmentation result of the right ventricle.

[0132] Furthermore, this disclosure proposes a cardiopulmonary analysis method, characterized by comprising: obtaining multiple right ventricular images using the aforementioned right ventricular segmentation method; determining the gender and chronic obstructive pulmonary disease (COPD) grade corresponding to the multiple right ventricular images; grouping the right ventricular images based on the gender and COPD grade; determining multiple morphological information corresponding to the grouped multiple right ventricular images; analyzing the relationship between different genders and COPD grades based on the multiple morphological information; and / or acquiring pulmonary function data and a preset classification model corresponding to the multiple right ventricular images; determining multiple morphological information corresponding to the multiple right ventricular images; and using the preset classification model, based on the multiple morphological information and the pulmonary function data, identifying dyspnea.

[0133] In the embodiments of this disclosure and other possible embodiments, the preset recognition model may be one or more of the following: Support Vector Machine (SVM), Multilayer Perceptron (MLP), Random Forest (RF), K Nearest Neighbors (KNN), Logistic Regression (LR), Decision Tree (DT), Gradient Boosting (GB), Linear Discriminant Analysis (LDA).

[0134] In embodiments of this disclosure and other possible embodiments, the method for identifying dyspnea based on the preset classification model, the multiple morphological information, and the lung function data includes: obtaining the gender of the patient to be identified, and obtaining morphological information affecting the gender and chronic obstructive pulmonary disease (COPD) grading; and using the preset classification model to identify dyspnea based on the morphological information affecting the gender and COPD grading and the lung function data. Specifically, before obtaining the morphological information affecting the gender and COPD grading, the morphological information affecting different genders and COPD gradings is determined, as detailed in the method for analyzing the relationship between different genders and COPD gradings based on the multiple morphological information.

[0135] In the embodiments of this disclosure and other possible embodiments, the method for identifying dyspnea using the preset classification model based on the morphological information affecting the gender and chronic obstructive pulmonary disease (COPD) grading and the lung function data includes: selecting corresponding preset local features from the morphological information and the lung function data respectively, determining global features based on the morphological information and the lung function data respectively, and fusing the preset local features and global features to obtain fused features; and using the fused features based on the preset identification model to complete the identification of the patient's dyspnea.

[0136] In embodiments of this disclosure, before acquiring the lung function data corresponding to the patient to be analyzed, the method for determining the lung function data includes: performing a lung function test on the patient using a pulmonary function testing instrument to obtain the lung function data. In embodiments of this disclosure and other possible embodiments, the pulmonary function testing instrument is a medical device used to measure the volume of air inhaled and exhaled by the lungs, capable of performing lung function tests and tracking lung health. The lung function data may include one or more commonly used lung function testing parameters such as FVC, FEV1, and FEV1 / FVC.

[0137] In the embodiments of this disclosure and other possible embodiments, since local features are already set, the corresponding set local features can be selected from the morphological information and the lung function data. However, global features need to be further determined based on the morphological information and the lung function data. For example, if the set local features corresponding to the lung function data feature af are features a, c, and z, then features a, c, and z can be selected simply from the lung function data feature af.

[0138] In embodiments of this disclosure and other possible embodiments, the method for determining the set local features before selecting corresponding set local features from the morphological information and the lung function data includes: obtaining a set feature selection model; selecting corresponding set local features from multiple sets of morphological information and the lung function data based on the feature selection model; or, obtaining a set feature selection rule; selecting corresponding set local features from multiple sets of morphological information and the lung function data based on the feature selection rule. Wherein, the multiple sets of morphological information and the lung function data used to determine the set local features are the morphological information and the lung function data corresponding to multiple dyspnea identification labels when constructing the model (training the classifier). Similarly, constructing the model (training the classifier) ​​also requires global features; in this case, the multiple sets of morphological information and the lung function data used for fusion of global features are the morphological information and the lung function data corresponding to multiple dyspnea identification labels when constructing the model (training the classifier). Wherein, "multiple" refers to the number of training sets.

[0139] In embodiments of this disclosure and other possible embodiments, the method for determining global features based on the morphological information and the lung function data includes: acquiring a set feature fusion model; and globally fusing the morphological information and the lung function data based on the feature fusion model to obtain corresponding global features. Similarly, when constructing a model (training a classifier), the method for determining global features based on the morphological information and the lung function data includes: acquiring a set feature fusion model; and globally fusing the plurality of morphological information and lung function data based on the feature fusion model to obtain corresponding global features for training. For detailed implementation processes, please refer to the description of the method for determining global features based on the morphological information and the lung function data.

[0140] For example, in embodiments of this disclosure and other possible embodiments, the set feature selection model can be the Lasso model. Using the Lasso model, based on the dyspnea recognition corresponding to the lung image (0 - no dyspnea, 1 - dyspnea), the morphological information and the lung function data are selected to obtain the selected set morphological information and the lung function data (set local features). More specifically, the dyspnea grading as a label can be obtained according to the Improved Medicine Research Council (MMRC) scale survey, grading dyspnea into levels 0-4; the Lasso model can be implemented using the standard Python package LassoCV (Python 3.6).

[0141] In embodiments of this disclosure and other possible embodiments, the method for determining global features based on the morphological information and the lung function data includes: acquiring a set feature fusion model; and globally fusing the morphological information and the lung function data based on the feature fusion model to obtain corresponding global features. In embodiments of this disclosure and other possible embodiments, the feature fusion model may be a principal component analysis (PCA) model or a neural network.

[0142] Meanwhile, when constructing the model (training the classifier, i.e., the preset recognition model), the method for determining global features based on the morphological information and the lung function data includes: obtaining a set feature fusion model; and globally fusing the morphological information and the lung function data based on the feature fusion model to obtain corresponding global features for training.

[0143] In another embodiment of this disclosure, the method for globally fusing the morphological information and the lung function data to obtain corresponding global features includes: constructing a neural network; training the neural network using the morphological information and the lung function data; and globally fusing the morphological information and the lung function data based on the trained neural network to obtain corresponding global features. The neural network includes at least an input layer, a hidden layer, and an output layer; the parameters corresponding to the input layer, hidden layer, and output layer of the neural network are trained using the morphological information and the lung function data to obtain a trained neural network.

[0144] In embodiments of this disclosure, the method for fusing the set local features and global features to obtain the fused features includes: splicing the set local features and global features to obtain the fused features.

[0145] In the embodiments of this disclosure and other possible embodiments, the right ventricular images are grouped based on the gender and the grade of chronic obstructive pulmonary disease (COPD): for example, the male group includes 4 subgroups, each subgroup corresponding to grade 1-4 of COPD; the female group also includes 4 subgroups, each subgroup corresponding to grade 1-4 of COPD.

[0146] In an embodiment of this disclosure, the method for determining the chronic obstructive pulmonary disease (COPD) grading corresponding to the multiple right ventricular images includes: acquiring chest images corresponding to the multiple right ventricular images and a preset grading model; and using the preset grading model to grade the chest images corresponding to the multiple right ventricular images to determine the COPD grading corresponding to the multiple right ventricular images.

[0147] In embodiments of this disclosure and other possible embodiments, the classification of the patient for chronic obstructive pulmonary disease (COPD) based on the lung function data may be performed according to the Global Initiative for Chronic Obstructive Pulmonary Disease (GOLD) criteria accepted by the American Thoracic Society and the European Respiratory Society.

[0148] In the embodiments of this disclosure and other possible embodiments, the preset hierarchical model may be one or more of the following: Support Vector Machine (SVM), Multilayer Perceptron (MLP), Random Forest (RF), K Nearest Neighbors (KNN), Logistic Regression (LR), Decision Tree (DT), Gradient Boosting (GB), Linear Discriminant Analysis (LDA).

[0149] In embodiments of this disclosure and other possible embodiments, the method for classifying chest images corresponding to multiple right ventricular images using the preset classification model includes: segmenting the chest images into lung regions to obtain lung parenchyma images; extracting omics features and convolutional features from the lung parenchyma images respectively; and performing feature selection on the omics features and convolutional features based on COPD identification or classification labels corresponding to the chest images respectively to obtain selected radiomics features and selected convolutional features; performing a first fusion operation on the selected radiomics features and the selected convolutional features to obtain a first fusion feature; and classifying COPD based on the first fusion feature and the preset classification model. This disclosure extracts omics features and convolutional features from the lung parenchyma image, respectively. Based on the COPD grading label corresponding to the lung image, feature selection is performed on the omics features and convolutional features to obtain selected radiomics features and selected convolutional features. A first fusion operation is performed on the selected radiomics features and selected convolutional features to obtain a first fused feature. Based on the first fused feature and a preset classifier (preset grading model), COPD is graded. Compared with traditional methods, this disclosure adds convolutional features corresponding to the lung parenchyma image. By fusing radiomics features and convolutional features, the accuracy of COPD grading and the classification performance of the preset grading model are improved.

[0150] In the embodiments of this disclosure and other possible embodiments, the extraction of omics features from the lung parenchyma image can be achieved through a preset radiomics feature extraction model. The preset radiomics feature extraction model is an existing radiomics computation model, which can be obtained from the website https: / / pyradiomics.readthedocs.io / en / latest / index.html. The preset radiomics feature extraction model will not be described in detail here.

[0151] In the embodiments of this disclosure and other possible embodiments, the convolutional features of the lung parenchyma image can be extracted using a pre-trained feature extraction model. For example, the feature extraction model includes multiple convolutional layers, which are used to extract the convolutional features of the lung parenchyma image. Alternatively, the convolutional features of the lung parenchyma image can be extracted using transfer learning. The transfer learning model can be the segmentation model proposed in the paper Med3d: Transfer learning for 3D medical image analysis (Chen, S., K. Ma and Y. Zheng), where only the encoding structure 3D ResNet10, 3D ResNet18, or 3D ResNet34 is needed to extract the convolutional features of the lung parenchyma image.

[0152] Meanwhile, in this disclosure, the method of performing feature selection on the omics features and the convolutional features based on the COPD grading labels corresponding to the lung images to obtain selected radiomics features and selected convolutional features includes: obtaining a radiomics selection model; and, based on the radiomics selection model, performing feature selection on the omics features and the convolutional features by establishing the relationship between the COPD identification and / or grading labels and the corresponding omics features and the convolutional features to obtain selected radiomics features and selected convolutional features.

[0153] In the embodiments of this disclosure and other possible embodiments, the feature selection model may be a Lasso model. Using the Lasso model, feature selection is performed on the omics features and the convolutional features based on the COPD grading labels corresponding to the lung images, respectively, to obtain selected radiomics features and selected convolutional features. For example, if the number of omics features extracted from the lung parenchyma image is 1316, and feature selection is performed on the omics features and the convolutional features based on the COPD grading labels corresponding to the lung images, selected radiomics features are obtained. Then, a first fusion operation is performed on the selected radiomics features and the selected convolutional features to obtain a first fused feature; and based on the first fused feature and a preset classifier (preset grading model), COPD is graded. Simultaneously, the method of performing the first fusion operation on the selected radiomics features and the selected convolutional features to obtain the first fused feature includes: vector concatenation of the selected radiomics features and the selected convolutional features to obtain the first fused feature. For example, the selected radiomics features corresponding to the lung image to be processed (the chest image corresponding to the multiple right ventricular images) are [1, 2, 4], where 1, 2, 4 correspond to radiomics features with different names. The selected convolutional feature corresponding to this lung image to be processed is [0, 2, 1]. Then, the selected radiomics features and the selected convolutional features are vector concatenated to obtain the first fusion feature as [1, 2, 4, 0, 2, 1] or [0, 2, 1, 1, 2, 4].

[0154] In embodiments of this disclosure, the method for analyzing the relationship between different genders and chronic obstructive pulmonary disease (COPD) grades based on the plurality of morphological information includes: obtaining a set significance value; performing significance analysis on the plurality of morphological information corresponding to the different genders and COPD grades respectively to obtain a plurality of significance values; selecting morphological information corresponding to the set significance value that is less than or equal to the plurality of significance values ​​to obtain morphological information affecting different genders and COPD grades. The set significance value can be configured to 0.05, and those skilled in the art can configure the set significance value according to actual needs.

[0155] In embodiments of this disclosure and other possible embodiments, a method for determining multiple morphological information corresponding to the grouped multiple right ventricular images includes: acquiring a morphological calculation model and configuring parameters of the morphological calculation model; and determining multiple morphological information corresponding to the grouped multiple right ventricular images based on the morphological calculation model with configured parameters.

[0156] In the embodiments and other possible embodiments disclosed herein, the morphological calculation model is a PyRadiomics model, configured with the following parameters: Original (no filter applied). Multiple morphological information parameters can be 2D and / or 3D morphological information, such as: Mesh Volume, Voxel Volume, Surface Area, Surface Area to Volume ratio, Compactness, Spherical Disproportion, Maximum 3D Diameter, Maximum 2D Diameter, Major Axis Length, Minor Axis Length, Least Axis Length, Elongation, Flatness, Sphericity, etc. For details, please refer to https: / / pyradiomics.readthedocs.io / en / latest / .

[0157] In the embodiments of this disclosure and other possible embodiments, a method for obtaining multiple significance values ​​by performing significance analysis on the multiple morphological information corresponding to the different genders and chronic obstructive pulmonary disease grades includes: calculating multiple first significance values ​​of the multiple morphological information between each pair of subgroups in the four subgroups of the male group; and calculating multiple second significance values ​​of the multiple morphological information between each pair of subgroups in the four subgroups of the female group.

[0158] In embodiments of this disclosure and other possible embodiments, a method for selecting morphological information corresponding to a predetermined significance value that is less than or equal to a plurality of significance values ​​to obtain morphological information affecting different genders and chronic obstructive pulmonary disease (COPD) grading includes: in the male group, selecting morphological information corresponding to a predetermined significance value that is less than or equal to a plurality of first significance values ​​to obtain morphological information affecting COPD grading in males; and in the female group, selecting morphological information corresponding to a predetermined significance value that is less than or equal to a plurality of first significance values ​​to obtain morphological information affecting COPD grading in females.

[0159] In embodiments of this disclosure, the method further includes: acquiring the image features corresponding to the lung images in the chest images corresponding to the plurality of right ventricular images; and using the preset recognition model to identify dyspnea based on the plurality of morphological information, the lung function data, and the image features.

[0160] In the embodiments of this disclosure and other possible embodiments, the acquisition of the image features corresponding to the lung images in the chest images corresponding to the multiple right ventricular images may include the above-mentioned omics features and convolutional features. That is, the chest images are segmented into lung regions to obtain lung parenchyma images; and the omics features and convolutional features of the lung parenchyma images are extracted respectively. The extraction methods of the omics features and convolutional features are detailed in the description above.

[0161] In embodiments of this disclosure and other possible embodiments, the method for identifying dyspnea based on the plurality of morphological information, the lung function data, and the imaging features includes: selecting corresponding preset local features from the morphological information, the lung function data, and the imaging features respectively, determining global features based on the morphological information, the lung function data, and the imaging features respectively, and fusing the preset local features and the global features to obtain fused features; and using the fused features based on a preset recognition model to complete the identification of the patient's dyspnea.

[0162] In the embodiments of this disclosure and other possible embodiments, since local features are already defined, the corresponding defined local features can be selected from the morphological information and the lung function data. However, global features need to be further determined based on the morphological information and the lung function data. For example, if the defined local features corresponding to the lung function data feature af are features a, c, and z, then simply selecting features a, c, and z from the lung function data feature af is sufficient.

[0163] In embodiments of this disclosure and other possible embodiments, the method for determining the set local features before selecting corresponding set local features from the morphological information, the lung function data, and the image features includes: obtaining a set feature selection model; selecting corresponding set local features from multiple sets of morphological information, the lung function data, and the image features based on the feature selection model; or, obtaining a set feature selection rule; selecting corresponding set local features from multiple sets of morphological information, the lung function data, and the image features based on the feature selection rule. Wherein, the multiple sets of morphological information and the lung function data used to determine the set local features are the morphological information, the lung function data, and the image features corresponding to multiple dyspnea identification labels when constructing the model (training the classifier, i.e., the preset recognition model). Similarly, constructing the model (training the classifier) ​​also requires global features. In this case, the multiple sets of morphological information, the lung function data, and the image features used for fusion of global features are the morphological information, the lung function data, and the image features corresponding to multiple dyspnea identification labels when constructing the model (training the classifier). Wherein, "multiple" refers to the number of training sets.

[0164] In embodiments of this disclosure and other possible embodiments, the method for determining global features based on the morphological information, the lung function data, and the imaging features includes: acquiring a set feature fusion model; and globally fusing the morphological information, the lung function data, and the imaging features based on the feature fusion model to obtain corresponding global features. Similarly, when constructing a model (training a classifier), the method for determining global features based on the morphological information, the lung function data, and the imaging features includes: acquiring a set feature fusion model; and globally fusing the plurality of morphological information, the lung function data, and the imaging features based on the feature fusion model to obtain corresponding global features for training. For detailed implementation processes, please refer to the description of the method for determining global features based on the morphological information, the lung function data, and the imaging features.

[0165] For example, in embodiments of this disclosure and other possible embodiments, the set feature selection model can be the Lasso model. Using the Lasso model, based on the dyspnea recognition corresponding to the lung image (0 - no dyspnea, 1 - dyspnea), the morphological information, the lung function data, and the image features are selected to obtain the set set morphological information, the lung function data, and the image features (set local features).

[0166] In embodiments of this disclosure and other possible embodiments, the method for determining global features based on the morphological information, the lung function data, and the imaging features includes: acquiring a set feature fusion model; and globally fusing the morphological information, the lung function data, and the imaging features based on the feature fusion model to obtain corresponding global features. In embodiments of this disclosure and other possible embodiments, the feature fusion model may be a principal component analysis (PCA) model or a neural network.

[0167] Meanwhile, when constructing the model (training the classifier, i.e., the preset recognition model), the method for determining global features based on the morphological information, the lung function data, and the image features includes: obtaining a set feature fusion model; and globally fusing the morphological information, the lung function data, and the image features based on the feature fusion model to obtain corresponding global features for training.

[0168] In another embodiment of this disclosure, the method for globally fusing the morphological information, the lung function data, and the image features to obtain corresponding global features includes: constructing a neural network; training the neural network using the morphological information, the lung function data, and the image features; and globally fusing the morphological information, the lung function data, and the image features based on the trained neural network to obtain corresponding global features. The neural network includes at least an input layer, a hidden layer, and an output layer; the parameters corresponding to the input layer, hidden layer, and output layer of the neural network are trained using the morphological information and the lung function data to obtain a trained neural network.

[0169] In embodiments of this disclosure, the method for fusing the set local features and global features to obtain the fused features includes: splicing the set local features and global features to obtain the fused features.

[0170] The execution entity of the right ventricular segmentation and / or cardiopulmonary analysis method can be a right ventricular segmentation and / or cardiopulmonary analysis device. For example, the right ventricular segmentation and / or cardiopulmonary analysis method can be executed by a terminal device, server, or other processing device. The terminal device can be a user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, in-vehicle device, wearable device, etc. In some possible implementations, the right ventricular segmentation and / or cardiopulmonary analysis method can be implemented by a processor calling computer-readable instructions stored in memory.

[0171] Those skilled in the art will understand that, in the above-described method of the specific implementation, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0172] In addition, this disclosure also proposes a right ventricular segmentation device, which includes: a first acquisition unit for acquiring a chest image to be segmented, a first preset segmentation model, and a second preset segmentation model; a first segmentation unit for using the first preset segmentation model to perform overall segmentation of the chest image to be segmented into the left and right ventricles to obtain a first segmented image; a second segmentation unit for using the second preset segmentation model to segment the chest image to be segmented into the left ventricle to obtain a second segmented image; and a first determination unit for determining a right ventricular image based on the first segmented image and the second segmented image.

[0173] Furthermore, this disclosure also proposes a cardiopulmonary analysis device, comprising: a second determining unit for determining the gender and chronic obstructive pulmonary disease (COPD) grade corresponding to the multiple right ventricular images; a grouping unit for grouping the right ventricular images based on the gender and COPD grade; a third determining unit for determining multiple morphological information corresponding to the grouped right ventricular images; an analysis unit for analyzing the relationship between different genders and COPD grades based on the multiple morphological information; and / or, a second acquiring unit for acquiring pulmonary function data and a preset classification model corresponding to the multiple right ventricular images; a fourth determining unit for determining multiple morphological information corresponding to the multiple right ventricular images; and a recognition unit for recognizing dyspnea using the preset classification model based on the multiple morphological information and the pulmonary function data.

[0174] In some embodiments, the functions or modules of the apparatus provided in this disclosure can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0175] This disclosure also proposes a computer-readable storage medium storing computer program instructions that, when executed by a processor, implement the above-described method. The computer-readable storage medium may be a non-volatile computer-readable storage medium.

[0176] This disclosure also proposes an electronic device, including: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured as described above. The electronic device may be provided as a terminal, a server, or other type of device.

[0177] Figure 3 This is a block diagram illustrating an electronic device 800 according to an exemplary embodiment. For example, the electronic device 800 may be a mobile phone, computer, digital broadcasting terminal, messaging device, game console, tablet device, medical device, fitness equipment, personal digital assistant, or other terminal.

[0178] Reference Figure 3 The electronic device 800 may include one or more of the following components: a processing component 802, a memory 804, a power supply component 806, a multimedia component 808, an audio component 810, an input / output (I / O) interface 812, a sensor component 814, and a communication component 816.

[0179] Processing component 802 typically controls the overall operation of electronic device 800, such as operations associated with display, telephone calls, data communication, camera operation, and recording operations. Processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the methods described above. Furthermore, processing component 802 may include one or more modules to facilitate interaction between processing component 802 and other components. For example, processing component 802 may include a multimedia module to facilitate interaction between multimedia component 808 and processing component 802.

[0180] Memory 804 is configured to store various types of data to support the operation of electronic device 800. Examples of this data include instructions for any application or method operating on electronic device 800, contact data, phonebook data, messages, pictures, videos, etc. Memory 804 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0181] Power supply component 806 provides power to various components of electronic device 800. Power supply component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to electronic device 800.

[0182] Multimedia component 808 includes a screen that provides an output interface between the electronic device 800 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may sense not only the boundaries of the touch or swipe action but also the duration and pressure associated with the touch or swipe operation. In some embodiments, multimedia component 808 includes a front-facing camera and / or a rear-facing camera. When the electronic device 800 is in an operating mode, such as a shooting mode or a video mode, the front-facing camera and / or the rear-facing camera may receive external multimedia data. Each front-facing camera and rear-facing camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

[0183] Audio component 810 is configured to output and / or input audio signals. For example, audio component 810 includes a microphone (MIC) configured to receive external audio signals when electronic device 800 is in an operating mode, such as call mode, recording mode, and voice recognition mode. The received audio signals may be further stored in memory 804 or transmitted via communication component 816. In some embodiments, audio component 810 also includes a speaker for outputting audio signals.

[0184] I / O interface 812 provides an interface between processing component 802 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to, home buttons, volume buttons, power buttons, and lock buttons.

[0185] Sensor assembly 814 includes one or more sensors for providing state assessments of various aspects of electronic device 800. For example, sensor assembly 814 can detect the on / off state of electronic device 800, the relative positioning of components such as the display and keypad of electronic device 800, changes in position of electronic device 800 or a component of electronic device 800, the presence or absence of user contact with electronic device 800, orientation or acceleration / deceleration of electronic device 800, and temperature changes of electronic device 800. Sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, sensor assembly 814 may also include an accelerometer, gyroscope, magnetometer, pressure sensor, or temperature sensor.

[0186] Communication component 816 is configured to facilitate wired or wireless communication between electronic device 800 and other devices. Electronic device 800 can access wireless networks based on communication standards, such as WiFi, 2G, or 3G, or combinations thereof. In one exemplary embodiment, communication component 816 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, communication component 816 also includes a near-field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0187] In an exemplary embodiment, the electronic device 800 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the methods described above.

[0188] In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 804 including computer program instructions that can be executed by a processor 820 of an electronic device 800 to perform the above-described method.

[0189] Figure 4 This is a block diagram illustrating an electronic device 1900 according to an exemplary embodiment. For example, the electronic device 1900 may be provided as a server. (Refer to...) Figure 4 The electronic device 1900 includes a processing component 1922, which further includes one or more processors, and memory resources represented by memory 1932 for storing instructions, such as application programs, that can be executed by the processing component 1922. The application programs stored in memory 1932 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 1922 is configured to execute instructions to perform the methods described above.

[0190] Electronic device 1900 may also include a power supply component 1926 configured to perform power management of electronic device 1900, a wired or wireless network interface 1950 configured to connect electronic device 1900 to a network, and an input / output (I / O) interface 1958. Electronic device 1900 can operate on an operating system stored in memory 1932, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, or similar.

[0191] In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 1932 including computer program instructions that can be executed by a processing component 1922 of an electronic device 1900 to perform the above-described method.

[0192] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0193] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination thereof. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0194] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0195] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0196] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0197] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0198] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0199] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0200] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for right ventricular segmentation, characterized in that, include: Obtain the chest image to be segmented, the first preset segmentation model, and the second preset segmentation model; Using the first preset segmentation model, the chest image to be segmented is divided into the left and right ventricles to obtain the first segmented image; Using the second preset segmentation model, the left ventricle of the chest image to be segmented is segmented to obtain a second segmented image; Determining a right ventricular image based on the first segmented image and the second segmented image includes: subtracting the second segmented image from the first segmented image to determine the right ventricular image; Before acquiring the first preset segmentation model and the second preset segmentation model, the first preset segmentation model and the second preset segmentation model are trained respectively, including: acquiring an enhanced chest image and a corresponding unenhanced chest image, a first label image of the left ventricle and the right ventricle, and a second label image of the left ventricle for training; registering the enhanced chest image and the corresponding unenhanced chest image for training to obtain a registered image corresponding to the unenhanced chest image; performing cardiac region segmentation on the registered image to obtain a cardiac region image; performing three-dimensional reconstruction on the cardiac region image and the first label image respectively to obtain a first three-dimensional cardiac image and a first three-dimensional label image; projecting the first three-dimensional cardiac image and the first three-dimensional label image according to multiple set angles to obtain multiple second three-dimensional cardiac projection images and corresponding multiple The first preset segmentation model is trained using the heart region image, the first label image, the plurality of second three-dimensional heart projection images, and the corresponding plurality of second three-dimensional label projection images to obtain the first preset segmentation model; the heart region image and the second label image are reconstructed in three dimensions to obtain a third three-dimensional heart image and a third three-dimensional label image; the third three-dimensional heart image and the third three-dimensional label image are projected at multiple set angles to obtain a plurality of fourth three-dimensional heart projection images and the corresponding plurality of fourth three-dimensional label projection images; the second preset segmentation model is trained using the heart region image, the second label image, the plurality of fourth three-dimensional heart projection images, and the corresponding plurality of fourth three-dimensional label projection images to obtain the second preset segmentation model.

2. The right ventricular segmentation method according to claim 1, characterized in that, include: The chest image to be segmented is an unenhanced chest image.

3. A method for right ventricular segmentation, characterized in that, include: Obtain the chest image to be segmented, the first preset segmentation model, and the second preset segmentation model; Using the first preset segmentation model, the chest image to be segmented is divided into the left and right ventricles to obtain the first segmented image; Using the second preset segmentation model, the left ventricle of the chest image to be segmented is segmented to obtain a second segmented image; Determining a right ventricular image based on the first segmented image and the second segmented image includes: subtracting the second segmented image from the first segmented image to determine the right ventricular image; Before acquiring the first and second preset segmentation models, the first and second preset segmentation models are trained, including: acquiring enhanced chest images and corresponding non-enhanced chest images, first label images of the left and right ventricles, and a second label image of the left ventricle for training; performing cardiac region segmentation on the enhanced chest images and their corresponding non-enhanced chest images to obtain a third and a fourth segmentation image; registering the third and fourth segmentation images to obtain a cardiac registration image; performing 3D reconstruction on the cardiac registration image and the first label image to obtain a first 3D cardiac registration image and a first 3D label image; and projecting the first 3D cardiac registration image and the first 3D label image from multiple preset angles to obtain multiple second 3D cardiac projection registration images and... The process involves obtaining multiple second three-dimensional label projection images; training a first preset segmentation model using the registered heart image, the first label image, the multiple second three-dimensional registered heart projection images, and the corresponding multiple second three-dimensional label projection images; performing three-dimensional reconstruction on the registered heart image and the second label image to obtain a third three-dimensional registered heart image and a third three-dimensional label image; projecting the third three-dimensional registered heart image and the third three-dimensional label image from multiple set angles to obtain multiple fourth three-dimensional registered heart projection images and the corresponding multiple fourth three-dimensional label projection images; and training a second preset segmentation model using the registered heart image, the second label image, the multiple fourth three-dimensional registered heart projection images, and the corresponding multiple fourth three-dimensional label projection images to obtain a second preset segmentation model.

4. The right ventricular segmentation method according to claim 3, characterized in that, The chest image to be segmented is an unenhanced chest image.

5. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the right ventricular segmentation method according to any one of claims 1 to 4, thereby obtaining multiple right ventricular images.

6. The electronic device according to claim 5, characterized in that, The processor is configured to invoke instructions stored in the memory to execute: The gender and chronic obstructive pulmonary disease (COPD) grade corresponding to the multiple right ventricular images were determined respectively; Based on the gender and the classification of chronic obstructive pulmonary disease, the images of the right ventricle were grouped. Each of the following morphological information corresponding to the grouped right ventricular images is determined: grid volume, voxel volume, surface area, surface area to volume ratio, compactness, spherical disproportion, maximum three-dimensional diameter, maximum two-dimensional diameter, major axis length, minor axis length, minimum axis length, elongation, flatness, and sphericity. Based on the aforementioned morphological information, the relationship between different genders and chronic obstructive pulmonary disease (COPD) grading was analyzed.

7. The electronic device according to claim 6, characterized in that, Determining the chronic obstructive pulmonary disease (COPD) classification corresponding to the multiple right ventricular images includes: Obtain chest images corresponding to the multiple right ventricular images and a preset grading model; Using the preset grading model, the chest images corresponding to the multiple right ventricular images are graded to determine the chronic obstructive pulmonary disease (COPD) grade corresponding to the multiple right ventricular images.

8. The electronic device according to claim 6, characterized in that, The analysis of the relationship between different genders and chronic obstructive pulmonary disease (COPD) grades based on the aforementioned morphological information includes: Get the set significance value; Significance analysis was performed on the multiple morphological information corresponding to different genders and chronic obstructive pulmonary disease grades to obtain multiple significance values; By selecting morphological information that is less than or equal to the set significance value from among the multiple significance values, morphological information that affects different genders and chronic obstructive pulmonary disease grading is obtained.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the right ventricular segmentation method according to any one of claims 1 to 4, resulting in multiple right ventricular images.

10. The computer-readable storage medium according to claim 9, characterized in that, The computer program instructions are implemented when executed by the processor: Obtain lung function data and a preset classification model corresponding to the multiple right ventricular images; Each of the following morphological information corresponding to the multiple right ventricular images is determined: grid volume, voxel volume, surface area, surface area to volume ratio, compactness, spherical disproportion, maximum three-dimensional diameter, maximum two-dimensional diameter, major axis length, minor axis length, minimum axis length, elongation, flatness, and sphericity. Using the preset classification model, dyspnea is identified based on the multiple morphological information and the lung function data.

11. The computer-readable storage medium according to claim 10, characterized in that, When the computer program instructions are executed by the processor, they also achieve the following: Obtain the image features corresponding to the lung images in the chest images corresponding to the multiple right ventricular images; Using the preset classification model, dyspnea is identified based on the multiple morphological information, the lung function data, and the imaging features.

Images