Pulmonary arterial hypertension detection method fusing image features and tricuspid regurgitation velocity

Abstract

The invention provides a pulmonary arterial hypertension detection method fusing image features and tricuspid regurgitation velocity, which is used for creating a set of dual-model fusion architecture when pulmonary arterial hypertension detection is carried out by combining computed tomography images, B ultrasound tricuspid regurgitation velocity and neural network technologies. A series of optimal configuration schemes are continuously given from details, so that an efficient and high-precision pulmonary arterial hypertension detection effect can be realized, the workload and working errors required by related medical personnel can be effectively reduced in practical application, the labor cost is remarkably reduced, high-quality data support can be provided, and the accuracy of pulmonary arterial hypertension detection is improved. Diagnosis and treatment of pulmonary arterial hypertension can be better promoted, and good application prospects are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical treatment, in particular to a pulmonary hypertension detection method fusing image features and tricuspid regurgitation velocity. BACKGROUND

[0002] Computed Tomography (CT) images have high density resolution, that is, different tissues show good contrast in computed tomography images, which makes the display of pulmonary artery and its surrounding structures clearer, facilitating the identification of pulmonary artery dilation, pulmonary parenchyma changes, calcified lesions, and heart shape changes related to pulmonary hypertension, especially in evaluating the main trunk diameter of the pulmonary artery, right ventricular hypertrophy, and pericardial effusion. Computed tomography can provide more intuitive and quantitative data support. At present, conventional computed tomography cannot show the changes of pulmonary vascular pressure, but through indirect indicators such as the ratio of pulmonary artery diameter to descending aorta diameter (PA / Ao ratio), the possibility of pulmonary hypertension can still be preliminarily judged.

[0003] The multi-angle and multi-section imaging characteristics of computed tomography overcome the limitations of the overlapping of traditional X-ray images, making the observation of details such as pulmonary vascular pathways, bronchial structures, and lung texture changes more accurate. This makes computed tomography pulmonary artery imaging an important means of evaluating pulmonary hypertension. Through the combination of High Resolution Computed Tomography (HRCT) and Computed Tomography Angiography (CTA) technology, not only can the pulmonary artery course and branch clarity be observed, but also the assessment of whether the pulmonary parenchymal lesions are the underlying cause of pulmonary hypertension, such as pulmonary interstitial fibrosis and chronic thromboembolism, can be made.

[0004] Magnetic Resonance Imaging (MRI) has advantages in soft tissue contrast, and its effect in evaluating early changes in heart structure is better than that of computed tomography, especially in evaluating right heart function, but it is not as clear as computed tomography in displaying lung structure, and the image acquisition time is longer, which is limited by patient compliance and scanning conditions, making it difficult to be widely developed.

[0005] However, the spatial resolution of computed tomography images still has limitations compared to nuclear magnetic resonance imaging and X-ray, that is, the single pixel of the computed tomography image has limited expression ability, and it is difficult to accurately quantify the changes in small blood vessel diameter and small calcified lesions, thereby limiting the detection of early lesions and small anatomical abnormalities of pulmonary hypertension.

[0006] In addition, the computer tomography of pulmonary arterial hypertension shows many indirect signs, and there is no unified gold standard image feature, which makes the image-based judgment easy to be affected by human experience.

[0007] Therefore, under the background of widely introducing Artificial Intelligence (AI) technology in medical technology research and clinical work to realize the prediction of corresponding diseases, how to realize high-performance pulmonary arterial hypertension detection based on computer tomography images is still a major technical problem. SUMMARY

[0008] The present application provides a pulmonary arterial hypertension detection method fusing image features and tricuspid regurgitation velocity, which is used to create a double-model fusion architecture when combining computer tomography images, B-ultrasound tricuspid regurgitation velocity, and neural network technology to detect pulmonary arterial hypertension, and a series of optimization configuration schemes are further given in detail, thereby realizing efficient and high-precision pulmonary arterial hypertension detection effect, effectively reducing the workload and work error of relevant medical staff in actual application, significantly reducing labor costs, thereby providing high-quality data support, helping to better promote the diagnosis and treatment of pulmonary arterial hypertension, and having a better application prospect.

[0009] In a first aspect, the present application provides a multi-modal pulmonary arterial hypertension detection method fusing medical image features and tricuspid regurgitation velocity, the method comprising: obtaining a to-be-detected lung computer tomography image and a to-be-detected B-ultrasound tricuspid regurgitation velocity corresponding to a current pulmonary arterial hypertension detection task; inputting the to-be-detected lung computer tomography image and the to-be-detected B-ultrasound tricuspid regurgitation velocity into a pre-configured pulmonary arterial hypertension detection model, wherein the pulmonary arterial hypertension detection model comprises a tissue site detection model part and a pulmonary arterial hypertension detection model part, the tissue site detection model part is used to perform classification and identification of target tissue sites related to pulmonary arterial hypertension on the input lung computer tomography image, and the pulmonary arterial hypertension detection model part is used to perform pulmonary arterial hypertension detection processing on the classification and identification result output by the tissue site detection model part in combination with the input B-ultrasound tricuspid regurgitation velocity, and the obtained pulmonary arterial hypertension detection processing result is taken as the model output; extracting a target pulmonary arterial hypertension detection result output by the pulmonary arterial hypertension detection model.

[0010] In a second aspect, the present application provides a pulmonary arterial hypertension detection device fusing image features and tricuspid regurgitation velocity, the device comprising: an acquisition unit, configured to obtain a to-be-detected lung computer tomography image and a to-be-detected B-ultrasound tricuspid regurgitation velocity corresponding to a current pulmonary arterial hypertension detection task; The detection unit is configured to input the lung computed tomography image to be detected and the B-ultrasonic tricuspid regurgitation velocity to be detected into a preselected pulmonary arterial hypertension detection model, wherein the pulmonary arterial hypertension detection model comprises a tissue site detection model part and a pulmonary arterial hypertension detection model part, the tissue site detection model part is configured to perform classification and identification of a target tissue site related to a pulmonary arterial hypertension condition on the lung computed tomography image input into the model, and the pulmonary arterial hypertension detection model part is configured to perform pulmonary arterial hypertension detection processing on the classification and identification result output by the tissue site detection model part in combination with the B-ultrasonic tricuspid regurgitation velocity input into the model, and the obtained pulmonary arterial hypertension detection processing result is taken as the model output. The extraction unit is configured to extract the target pulmonary arterial hypertension detection result output by the pulmonary arterial hypertension detection model.

[0011] In a third aspect, the present application provides a processing device, comprising a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program in the memory to perform the method provided in the first aspect of the present application.

[0012] In a fourth aspect, the present application provides a computer readable storage medium, which stores a plurality of instructions, and the instructions are adapted to be loaded by a processor to execute the method provided in the first aspect of the present application.

[0013] From the above, the present application has the following beneficial effects: For the pulmonary arterial hypertension detection target of fusing image features and tricuspid regurgitation velocity, when the present application combines the computed tomography image, the B-ultrasonic tricuspid regurgitation velocity and the neural network technology to perform the pulmonary arterial hypertension detection, a double-model fusion architecture is created, and a series of optimization configuration schemes are further given in details, so that the efficient and high-precision pulmonary arterial hypertension detection effect can be realized, the workload and the work error of the related medical staff can be effectively reduced in the actual application, the artificial cost is significantly reduced, and therefore the high-quality data support can be provided, which is helpful to better promote the diagnosis and treatment of the pulmonary arterial hypertension, and has a better application prospect. BRIEF DESCRIPTION OF DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0015] Figure 1 A flowchart of the pulmonary arterial hypertension detection method of the present application fusing image features and tricuspid regurgitation velocity is shown in the figure. Figure 2This is a schematic diagram of a pulmonary hypertension detection device that integrates imaging features and tricuspid regurgitation velocity according to this application. Figure 3 This is a schematic diagram of one type of processing equipment used in this application. Detailed Implementation

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0017] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules is not necessarily limited to those explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices. The naming or numbering of steps appearing in this application does not imply that the steps in the method flow must be performed in the chronological / logical order indicated by the naming or numbering. The execution order of named or numbered process steps can be changed according to the desired technical purpose, as long as the same or similar technical effect is achieved.

[0018] The module division described in this application is a logical division. In practical applications, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the coupling or direct coupling or communication connection between modules shown or discussed may be through some interfaces, and the indirect coupling or communication connection between modules may be electrical or other similar forms, none of which are limited in this application. Furthermore, the modules or sub-modules described as separate components may or may not be physically separated, may or may not be physical modules, or may be distributed in multiple circuit modules. Some or all of the modules may be selected to achieve the purpose of the solution in this application according to actual needs.

[0019] Before introducing the pulmonary hypertension detection method based on the fusion of imaging features and tricuspid regurgitation velocity provided in this application, the background information involved in this application will be introduced first.

[0020] The pulmonary hypertension detection method, apparatus, and computer-readable storage medium provided in this application, which integrate imaging features and tricuspid regurgitation velocity, can be applied to processing equipment. They create a dual-model fusion architecture for pulmonary hypertension detection by combining computed tomography images and neural network technology, and further provide a series of optimized configuration schemes in detail. This enables efficient and high-precision pulmonary hypertension detection, effectively reducing the workload and errors required by medical personnel in practical applications, significantly lowering labor costs, and providing high-quality data support. This contributes to better advancing the diagnosis and treatment of pulmonary hypertension and has promising application prospects.

[0021] The pulmonary hypertension detection method fusing image features and tricuspid regurgitation velocity mentioned in this application can be implemented by a pulmonary hypertension detection device fusing image features and tricuspid regurgitation velocity, or by different types of processing devices such as servers, physical hosts, or user equipment (UE) integrating such a device. The pulmonary hypertension detection device fusing image features and tricuspid regurgitation velocity can be implemented in hardware or software. The UE can be a smartphone, tablet, laptop, desktop computer, or personal digital assistant (PDA) or other terminal device. The processing devices can be configured in a device cluster.

[0022] It is understandable that the solution in this application is mainly based on the data processing carried out on the basis of the already acquired or ready-made computed tomography images. Therefore, the processing equipment that implements the pulmonary hypertension detection method of fusing image features and tricuspid regurgitation velocity of this application, or the processing equipment that carries the corresponding application service of the pulmonary hypertension detection method of fusing image features and tricuspid regurgitation velocity of this application, usually only needs to meet the required data processing capabilities. The specific equipment type and equipment deployment form can be flexibly configured according to the actual situation.

[0023] In specific applications, if the acquisition and processing of specific computed tomography (CT) images are involved, i.e., CT scan processing, then the corresponding CT scan equipment can be included in the scope of the processing equipment of this application, or the CT scan service can be provided as an external linkage device.

[0024] Furthermore, if further processing is involved, such as displaying results or processing progress, it is obviously necessary to make further adaptive configurations for the equipment type and deployment form of the processing equipment based on the actual situation.

[0025] Taking the requirement to configure display functions as an example, the processing device can display specific content through its own configured display screen (including touch screen), external display devices, or other devices with display screens.

[0026] The following section introduces the method for detecting pulmonary hypertension by fusing imaging features and tricuspid regurgitation velocity provided in this application.

[0027] First, refer to Figure 1 , Figure 1 This paper illustrates a flowchart of a pulmonary hypertension detection method that integrates imaging features and tricuspid regurgitation velocity, as described in this application. The pulmonary hypertension detection method that integrates imaging features and tricuspid regurgitation velocity provided in this application may specifically include the following steps S101 to S103: Step S101: Obtain the computed tomography image of the lung to be detected and the tricuspid regurgitation velocity of the ultrasound corresponding to the current pulmonary hypertension detection task; As is easily understood, in the case of the novel pulmonary hypertension detection model designed based on artificial intelligence technology in this application, the specific application of the pulmonary hypertension detection model requires the acquisition of the current lung computed tomography image to be processed. For ease of explanation, the image to be processed is referred to as the lung computed tomography image to be detected (which can be simply referred to as the CT image to be detected) and the tricuspid regurgitation velocity (TR velocity) to be detected by ultrasound. The latter is an existing clinical indicator for assessing cardiopulmonary function.

[0028] Furthermore, it can be seen that in practical applications, the application processing of the pulmonary hypertension detection model in this application is usually initiated in the form of a schematic task.

[0029] The acquisition and processing of the computed tomography (CT) images of the lungs to be detected and the tricuspid regurgitation velocity to be detected by ultrasound can be achieved by directly extracting them from the task information of the current pulmonary hypertension detection task, extracting them according to the instructions and guidance of the task information of the current pulmonary hypertension detection task, or performing CT scan processing and ultrasound tricuspid regurgitation velocity examination processing in real time according to the instructions and guidance of the task information of the current pulmonary hypertension detection task (i.e., real-time acquisition and processing). The specific acquisition method is quite flexible and can be configured according to the actual situation.

[0030] Step S102: Input the lung computed tomography image to be detected and the tricuspid regurgitation velocity of the ultrasound to be detected into the pre-selected pulmonary hypertension detection model. The pulmonary hypertension detection model includes a tissue site detection model and a pulmonary hypertension detection model. The tissue site detection model is used to classify and identify target tissue sites related to pulmonary hypertension based on the lung computed tomography image input into the model. The pulmonary hypertension detection model is used to combine the classification and identification results output by the tissue site detection model with the tricuspid regurgitation velocity of the ultrasound input into the model to perform pulmonary hypertension detection processing. The resulting pulmonary hypertension detection processing result is used as the model output. It should be understood that the pulmonary hypertension detection model specifically designed in this application is divided into two parts: The first part is the tissue site detection model. The input to this tissue site detection model is the lung computed tomography image input to the model. It is used to classify and identify target tissue sites related to pulmonary hypertension based on the lung computed tomography image input to the model. The second part is the pulmonary hypertension detection model. The output of this pulmonary hypertension detection model is the output of the pulmonary hypertension detection model and the tricuspid regurgitation velocity input by the model (in some special cases, other types of clinical parameters may also be involved). The pulmonary hypertension detection model performs pulmonary hypertension detection processing by combining the classification and recognition results output by the tissue site detection model with the tricuspid regurgitation velocity input by the model.

[0031] These two types of models are typically built using different model architectures to correspond to different prediction targets, thereby achieving more suitable and refined processing results.

[0032] Regarding the former, namely the tissue site detection model, it is understandable that in reality, computed tomography (CT) images contain multiple volumetric data. A patient's CT scan often contains hundreds of slices, but the number of slices that actually contain effective diagnostic information is very limited. Therefore, in traditional processing methods, the screening and annotation process requires professional doctors and consumes a lot of time to complete, resulting in high annotation costs and difficulty in ensuring consistency. In addition, the spatial resolution limitation of CT scan images can easily lead to blurred lesion boundaries and decreased recognition accuracy, further affecting the performance of the AI ​​model. Under these circumstances, the goal of the tissue site detection model is to efficiently and accurately screen and identify target tissue sites that are highly relevant to the pulmonary hypertension detection logic designed in this application from the large number of slices contained in the lung CT scan image input to the model.

[0033] In layman's terms, computed tomography (CT) images are quite noisy and are usually composed of regions of interest (ROIs) and background regions. ROIs often contain key diagnostic information. Therefore, the tissue site detection model (i.e., the detection algorithm) detects a very small number of ROIs from a large number of previous images, and then the pulmonary hypertension detection model performs lesion analysis on them.

[0034] The latter, the pulmonary hypertension detection model, is based on the highly simplified ROI region provided by the tissue site detection model. It combines the tricuspid regurgitation velocity input by the B-mode ultrasound into the model, thereby efficiently utilizing computing resources to capture the features of the lesion image in a short time, and then accurately inferring and predicting the corresponding pulmonary hypertension.

[0035] Thus, through the aforementioned dual-model fusion architecture, or dual-layer AI inference mechanism, the goal of high-performance pulmonary hypertension detection can be achieved.

[0036] Step S103: Extract the target pulmonary hypertension detection results output by the pulmonary hypertension detection model.

[0037] It is understandable that after the pulmonary hypertension detection model has completed the corresponding pulmonary hypertension detection processing for the current lung computed tomography image and the tricuspid regurgitation velocity on ultrasound, it is obvious that the target pulmonary hypertension detection result output by the pulmonary hypertension detection model (specifically the pulmonary hypertension detection model part configured in the second layer of the pulmonary hypertension detection model) can be extracted, which corresponds to both the previous lung computed tomography image and the tricuspid regurgitation velocity on ultrasound.

[0038] At this point, further processing can be carried out according to the corresponding data application requirements.

[0039] For example, the results of the pulmonary hypertension test can be stored locally or remotely, displayed, and output as a notification of completion of the test, pushed to the system, or further analyzed and processed. The specific processing can be flexibly adjusted according to the pre-configured and real-time data application strategies to meet the diverse data application needs in actual situations and better assist the diagnosis and treatment of pulmonary hypertension from various dimensions.

[0040] Thus, from the above Figure 1As can be seen from the embodiments shown, for the detection target of pulmonary hypertension by fusing imaging features and tricuspid regurgitation velocity, this application has created a dual-model fusion architecture when combining computed tomography images, ultrasound tricuspid regurgitation velocity, and neural network technology for pulmonary hypertension detection. Furthermore, it provides a series of optimized configuration schemes in detail, thereby achieving efficient and high-precision pulmonary hypertension detection. In practical applications, this can effectively reduce the workload and errors required by relevant medical personnel, significantly reduce labor costs, and thus provide high-quality data support, helping to better advance the diagnosis and treatment of pulmonary hypertension, and has promising application prospects.

[0041] Continue with the above Figure 1 The steps of the illustrated embodiment and their possible implementation methods in practical applications are described in detail.

[0042] As an exemplary embodiment, the tissue site detection model part configured in the first layer of the pulmonary hypertension detection model specifically targets the ROI region, which may include the pulmonary artery, descending aorta, left ventricle, and right ventricle.

[0043] For these specific tissue sites, the pulmonary hypertension detection model configured in the second layer can specifically analyze the pathological image features related to pulmonary hypertension, such as pulmonary artery dilation (corresponding to the diameter of the main pulmonary artery), changes in lung parenchyma (including calcification lesions, pulmonary interstitial fibrosis, and chronic thromboembolism), changes in lung texture, and changes in heart morphology (including right ventricular hypertrophy and pericardial effusion), to further determine the specific pulmonary hypertension condition.

[0044] Furthermore, this application also provides a specific implementation scheme for the specific neural network architecture involved in the two-layer model part of the pulmonary hypertension detection model, starting from the details.

[0045] Specifically, as an exemplary embodiment, on the one hand, the tissue site detection model configured in the first layer of the pulmonary hypertension detection model specifically adopts the HydraNet model; on the other hand, the pulmonary hypertension detection model configured in the second layer of the pulmonary hypertension detection model specifically adopts the multimodal convolutional neural network-visual self-attention (CNN-ViT) model that integrates image features and tricuspid regurgitation velocity.

[0046] More specifically, as mentioned earlier, computed tomography (CT) images often contain much content that is not helpful in assessing pulmonary hypertension. They also suffer from low contrast and unclear edges, which greatly complicates lesion segmentation and disease diagnosis. Therefore, in the design of this application, we need to perform data augmentation and filtering on CT images to improve our accuracy.

[0047] This application specifically employs the HydraNet model to classify and identify computed tomography images related to pulmonary hypertension, accurately locating various tissue sites (pulmonary artery, descending aorta, left and right ventricles, etc.) from hundreds of computed tomography slices, and determining the corresponding ROI regions and background regions.

[0048] The HydraNet model is a multi-branch, multi-scale convolutional neural network architecture that can simultaneously extract information from different levels and scales in an image. It is suitable for medical image segmentation tasks with complex structures or blurred boundaries. It introduces multiple decoding branches on the basis of the backbone network, with each branch decoding features at different levels, thereby achieving more refined spatial information recovery. Its advantage lies in its ability to enhance the expression of features at different resolutions in the image and improve the model's ability to recognize small target structures (such as pulmonary arteries). It is particularly suitable for the accurate depiction of blood vessel boundaries and morphology in computed tomography images.

[0049] For the subsequent lesion analysis and processing in the ROI region, this application specifically used the multimodal CNN-ViT model. Compared with traditional convolutional neural network (CNN)-based detection methods, such as Faster R-CNN and YOLOv5, this image detection algorithm has significantly improved in complex structure recognition and global feature modeling.

[0050] Its main advantage lies in the fact that it does not rely on convolution operations of local receptive fields. Instead, it divides the image into fixed-size patches (i.e., a continuous time segment involving two key parameters: Patch Length and Stride), and feeds them into the Transformer module as input sequences after linear embedding. This structure makes the model more efficient in processing the relationship between distant pixels, and is especially suitable for detecting complex anatomical regions such as the pulmonary artery and ventricular wall in computed tomography images with low contrast, blurred boundaries, or severe structural overlap.

[0051] Thus, the pulmonary hypertension detection performance of the dual-model fusion architecture or dual-layer AI inference mechanism designed in this application is further enhanced through the two specific model architectures mentioned above.

[0052] The multimodal CNN-ViT model includes a CNN branch and a ViT branch. As an example, the CNN branch can specifically use a ResNet-50 network to extract local texture and spatial features. More specifically, the ResNet-50 network is responsible for extracting the deep visual features of the images output by the tissue site detection model (HydraNet model), i.e., the classification and recognition results of target tissue sites related to pulmonary hypertension. It then continues to model global dependencies through the ViT branch. Finally, the image representations output by the two branches and the standardized model input ultrasound tricuspid regurgitation velocity (Transformervelocity) are input into a multilayer perceptron (MLP) for final feature fusion and classification.

[0053] Furthermore, it is understood that before the pulmonary hypertension detection model is put into practical use, it may involve corresponding training processes. Therefore, as an exemplary embodiment, the method of this application may also include a corresponding model training step, namely: Obtain computed tomography (CT) images of the lungs of the sample and ultrasound images of the tricuspid regurgitation velocity of the sample; Configure appropriate annotations for the lung computed tomography images and the tricuspid regurgitation velocity of the sample via ultrasound. A pulmonary hypertension detection model was trained using labeled sample lung computed tomography images and sample B-mode ultrasound tricuspid regurgitation velocities.

[0054] The sample lung computed tomography (CT) images can be images acquired from real patients (involving real-time image acquisition and processing or extraction and processing of existing images), images obtained through further modifications based on real images, or even directly drawn and generated images, to meet diverse sample configuration needs and cover as many pulmonary hypertension conditions as possible, especially those rare in reality. Similarly, the sample B-ultrasound tricuspid regurgitation velocity corresponds to the sample lung CT images; more simply, both correspond to the same real or virtual patient.

[0055] The subsequent annotation process is usually done manually, or it can be done using an automated annotation tool. Before using the automated annotation tool, the corresponding automated annotation logic needs to be pre-configured.

[0056] Thus, under the corresponding model training scheme, based on the configured training samples, namely the labeled lung computed tomography images and the sample B-ultrasound tricuspid regurgitation velocity, specific model training can be carried out, and the trained model that can be put into practical use is referred to as the pulmonary hypertension detection model involved in this application.

[0057] The specific model training scheme and loss function used in the model training process can be understood as follows: it can be an existing scheme, a further optimization based on the existing scheme, or a novel scheme developed in-house, which can be configured according to actual needs.

[0058] As an example, the lung computed tomography images and tricuspid regurgitation velocities of the samples can be randomly divided into training and validation sets in a 3:1 ratio. The training set is used to train the model, and the validation set is used to evaluate the accuracy of the classification performance of the constructed model.

[0059] Furthermore, for the pulmonary hypertension detection model portion configured as the second layer in the pulmonary hypertension detection model, this application also provides a more practical implementation scheme.

[0060] Specifically, as an exemplary embodiment, the training process of the pulmonary hypertension detection model may include the following processing: For a pre-trained multimodal CNN-ViT model, pre-trained weights are loaded and fine-tuned using transfer learning.

[0061] It is understandable that combining pre-training mechanisms with fine-tuning mechanisms based on transfer learning can help achieve more convenient and efficient model training for the pulmonary hypertension detection model.

[0062] As an example, in the deep learning-based image detection model, this application specifically uses the PyTorch platform for model configuration and uses a multimodal CNN-ViT model pre-trained on the ImageNet dataset to configure the pulmonary hypertension detection model part of this application. During the retraining process, the model is initialized with pre-trained weights, and the multimodal CNN-ViT model is fine-tuned through transfer learning to adapt to the ROI region recognition task involved in this application. With the network architecture remaining unchanged, the training process stops when the loss function on the validation set no longer decreases.

[0063] Meanwhile, in terms of details, the operation of the pulmonary hypertension detection model, as an exemplary embodiment, may also include: Before inputting the classification and recognition results output by the tissue site detection model into the pulmonary hypertension detection model, the ROI region image corresponding to the classification and recognition results is subjected to contrast enhancement processing.

[0064] Understandably, this application inserts a contrast enhancement processing step between the tissue site detection model and the pulmonary hypertension detection model to strengthen the corresponding image edge features at the detail level. Then, the subsequent pulmonary hypertension detection model performs the specific processing to determine the presence of lesions more efficiently and accurately based on the image size and its characteristics.

[0065] Furthermore, this application also provides a practical implementation scheme for the subsequent result display stage. Correspondingly, as an exemplary embodiment, the method of this application may further include: The results of the target pulmonary hypertension detection are displayed. In addition to the pulmonary hypertension category identification results, the target pulmonary hypertension detection results also include the tricuspid regurgitation velocity of the ultrasound to be detected and the corresponding lesion image features. The lesion image features specifically involve the image features of pulmonary artery dilation, lung parenchymal changes, cardiac morphological changes or chronic thromboembolism.

[0066] As can be seen, when outputting whether pulmonary hypertension exists, if pulmonary hypertension is present, the pulmonary hypertension detection model provides the following results: in addition to the specific pulmonary hypertension category identification results (e.g., pulmonary hypertension caused by left heart disease, pulmonary hypertension caused by lung disease, chronic thromboembolic pulmonary hypertension), it can also provide the initial tricuspid regurgitation velocity on ultrasound and corresponding pathological features. This provides richer image content and better reference, allowing medical staff to understand the specific situation more intuitively and accurately, thus helping to advance the next step of diagnosis and treatment more accurately in the first instance.

[0067] In conclusion, this application, focusing on the detection target of pulmonary hypertension by fusing imaging features and tricuspid regurgitation velocity, addresses a series of issues related to data acquisition, standardized annotation, sample size expansion, and model robustness improvement from the perspective of specific clinical applications. This results in a high-performance pulmonary hypertension detection model, which will contribute significantly to advancing AI-based pulmonary hypertension identification from research to practical application.

[0068] The above is an introduction to the pulmonary hypertension detection method based on the fusion of imaging features and tricuspid regurgitation velocity provided in this application. To facilitate better implementation of the pulmonary hypertension detection method based on the fusion of imaging features and tricuspid regurgitation velocity provided in this application, this application also provides a pulmonary hypertension detection device based on the fusion of imaging features and tricuspid regurgitation velocity from the perspective of functional modules.

[0069] See Figure 2 , Figure 2 This is a schematic diagram of a pulmonary hypertension detection device that integrates imaging features and tricuspid regurgitation velocity, as described in this application. Specifically, the pulmonary hypertension detection device 200 that integrates imaging features and tricuspid regurgitation velocity may include the following structure: The acquisition unit 201 is used to acquire the lung computed tomography image to be detected and the tricuspid regurgitation velocity to be detected corresponding to the current pulmonary hypertension detection task. The detection unit 202 is used to input the lung computed tomography image to be detected and the tricuspid regurgitation velocity of the ultrasound to be detected into a pre-selected pulmonary hypertension detection model. The pulmonary hypertension detection model includes a tissue site detection model part and a pulmonary hypertension detection model part. The tissue site detection model part is used to classify and identify target tissue sites related to pulmonary hypertension based on the lung computed tomography image input into the model. The pulmonary hypertension detection model part is used to combine the classification and identification results output by the tissue site detection model part with the tricuspid regurgitation velocity of the ultrasound input into the model to perform pulmonary hypertension detection processing. The resulting pulmonary hypertension detection processing result is used as the model output. Extraction unit 203 is used to extract the target pulmonary hypertension detection results output by the pulmonary hypertension detection model.

[0070] In yet another exemplary embodiment, the target tissue site specifically includes the pulmonary artery, descending aorta, left ventricle, and right ventricle.

[0071] In yet another exemplary embodiment, the tissue site detection model specifically employs the HydraNet model, and the pulmonary hypertension detection model specifically employs the multimodal CNN-ViT model.

[0072] In another exemplary embodiment, the multimodal CNN-ViT model includes a CNN branch and a ViT branch. The CNN branch specifically adopts a ResNet-50 network and is responsible for extracting deep visual features of the classification and recognition results output by the tissue site detection model. It continues to model global dependencies through the ViT branch. Finally, the image representations output by the two branches and the standardized model input ultrasound tricuspid regurgitation velocity are input into a multilayer perceptron for final feature fusion and classification.

[0073] In yet another exemplary embodiment, the apparatus further includes a training unit 204 for: Obtain computed tomography (CT) images of the lungs of the sample and ultrasound images of the tricuspid regurgitation velocity of the sample; Configure appropriate annotations for the lung computed tomography images and the tricuspid regurgitation velocity of the sample via ultrasound. A pulmonary hypertension detection model was trained using labeled sample lung computed tomography images and sample B-mode ultrasound tricuspid regurgitation velocities.

[0074] In yet another exemplary embodiment, the operation of the pulmonary hypertension detection model further includes: Before inputting the classification and recognition results output by the tissue site detection model into the pulmonary hypertension detection model, the ROI region image corresponding to the classification and recognition results is subjected to contrast enhancement processing.

[0075] In yet another exemplary embodiment, the apparatus further includes a display unit 205 for: The results of the target pulmonary hypertension detection are displayed. In addition to the pulmonary hypertension category identification results, the target pulmonary hypertension detection results also include the tricuspid regurgitation velocity of the ultrasound to be detected and the corresponding lesion image features. The lesion image features specifically involve the image features of pulmonary artery dilation, lung parenchymal changes, cardiac morphological changes or chronic thromboembolism.

[0076] This application also provides a processing device from a hardware architecture perspective, see [link / reference]. Figure 3 , Figure 3 This diagram illustrates a structural schematic of the processing device of this application. Specifically, the processing device may include a processor 301, a memory 302, and an input / output device 303. The processor 301 executes the computer program stored in the memory 302 to implement, for example... Figure 1 The steps of the pulmonary hypertension detection method that fuses image features and tricuspid regurgitation velocity in the corresponding embodiment; or, when the processor 301 executes the computer program stored in the memory 302, it implements as follows: Figure 2 Corresponding to the functions of each unit in the embodiment, the memory 302 is used to store the functions executed by the processor 301 as described above. Figure 1 The computer program required for the pulmonary hypertension detection method that integrates imaging features and tricuspid regurgitation velocity in the corresponding embodiment.

[0077] For example, a computer program may be divided into one or more modules / units, one or more of which are stored in memory 302 and executed by processor 301 to complete this application. One or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in a computer device.

[0078] The processing device may include, but is not limited to, processor 301, memory 302, and input / output device 303. Those skilled in the art will understand that the illustrations are merely examples of the processing device and do not constitute a limitation on the processing device. It may include more or fewer components than illustrated, or combine certain components, or different components. For example, the processing device may also include network access devices, buses, etc., and processor 301, memory 302, input / output device 303, etc., are connected via a bus.

[0079] Processor 301 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the processing device, connecting various parts of the device through various interfaces and lines.

[0080] The memory 302 can be used to store computer programs and / or modules. The processor 301 implements various functions of the computer device by running or executing the computer programs and / or modules stored in the memory 302 and by calling data stored in the memory 302. The memory 302 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function, etc.; the data storage area may store data created according to the use of the processing device, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0081] When processor 301 executes a computer program stored in memory 302, it can specifically perform the following functions: Acquire the computed tomography (CT) image of the lungs to be detected and the tricuspid regurgitation velocity to be detected corresponding to the current pulmonary hypertension detection task; The computed tomography (CT) image of the lung to be detected and the tricuspid regurgitation velocity of the ultrasound to be detected are input into a pre-selected pulmonary hypertension detection model. The pulmonary hypertension detection model includes a tissue site detection model and a pulmonary hypertension detection model. The tissue site detection model is used to classify and identify target tissue sites related to pulmonary hypertension based on the CT image of the lung input to the model. The pulmonary hypertension detection model is used to process the classification and identification results output by the tissue site detection model and the tricuspid regurgitation velocity of the ultrasound input to the model to detect pulmonary hypertension. The resulting pulmonary hypertension detection processing result is used as the model output. Extract the target pulmonary hypertension detection results output by the pulmonary hypertension detection model.

[0082] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the pulmonary hypertension detection device, processing equipment, and its corresponding units that integrate imaging features and tricuspid regurgitation velocity described above can be found in reference to... Figure 1 The description of the pulmonary hypertension detection method that integrates imaging features and tricuspid regurgitation velocity in the corresponding embodiment will not be repeated here.

[0083] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0084] Therefore, this application provides a computer-readable storage medium storing a plurality of instructions that can be loaded by a processor to execute the present application. Figure 1 The steps of the pulmonary hypertension detection method that integrates imaging features and tricuspid regurgitation velocity in the corresponding embodiment can be referred to as follows: Figure 1 The description of the pulmonary hypertension detection method that integrates imaging features and tricuspid regurgitation velocity in the corresponding embodiments will not be repeated here.

[0085] The computer-readable storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0086] Because of the instructions stored in the computer-readable storage medium, the present application can be executed as described above. Figure 1 The steps of the pulmonary hypertension detection method that integrates imaging features and tricuspid regurgitation velocity in the corresponding embodiment can therefore achieve the results of this application. Figure 1The beneficial effects that the pulmonary hypertension detection method integrating imaging features and tricuspid regurgitation velocity can achieve in the corresponding embodiments are detailed in the preceding description and will not be repeated here.

[0087] The foregoing has provided a detailed description of the pulmonary hypertension detection method, apparatus, processing device, and computer-readable storage medium for fusing imaging features and tricuspid regurgitation velocity provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for detecting pulmonary hypertension by fusing imaging features and tricuspid regurgitation velocity, characterized in that, The method includes: Acquire the computed tomography (CT) image of the lungs to be detected and the tricuspid regurgitation velocity to be detected corresponding to the current pulmonary hypertension detection task; The computed tomography (CT) image of the lung to be detected and the tricuspid regurgitation velocity of the ultrasound to be detected are input into a pre-selected pulmonary hypertension detection model. The pulmonary hypertension detection model includes a tissue site detection model and a pulmonary hypertension detection model. The tissue site detection model is used to classify and identify target tissue sites related to pulmonary hypertension based on the CT image of the lung input to the model. The pulmonary hypertension detection model is used to combine the classification and identification results output by the tissue site detection model with the tricuspid regurgitation velocity of the ultrasound input to the model to perform pulmonary hypertension detection processing. The resulting pulmonary hypertension detection processing result is used as the model output. Extract the target pulmonary hypertension detection results output by the pulmonary hypertension detection model.

2. The method according to claim 1, characterized in that, The target tissue sites specifically include the pulmonary artery, descending aorta, left ventricle, and right ventricle.

3. The method according to claim 1, characterized in that, The tissue site detection model specifically adopts the HydraNet model, and the pulmonary hypertension detection model specifically adopts the multimodal CNN-ViT model.

4. The method according to claim 3, characterized in that, The multimodal CNN-ViT model includes a CNN branch and a ViT branch. The CNN branch specifically uses a ResNet-50 network and is responsible for extracting deep visual features from the classification and recognition results output by the tissue detection model. It also continues to model global dependencies through the ViT branch. Finally, the image representations output by the two branches and the standardized tricuspid regurgitation velocity input to the model via ultrasound are input into a multilayer perceptron for final feature fusion and classification.

5. The method according to claim 1, characterized in that, The method further includes: Obtain computed tomography (CT) images of the lungs of the sample and the corresponding ultrasound tricuspid regurgitation velocity of the sample; The lung computed tomography images of the sample and the tricuspid regurgitation velocity on ultrasound of the sample are assigned corresponding annotations; The pulmonary hypertension detection model was trained using labeled computed tomography images of the lungs of the sample and B-mode ultrasound images of the tricuspid regurgitation velocity of the sample.

6. The method according to claim 1, characterized in that, The operation of the pulmonary hypertension detection model also includes: Before inputting the classification and recognition results output by the tissue site detection model into the pulmonary hypertension detection model, the ROI region image corresponding to the classification and recognition results is subjected to contrast enhancement processing.

7. The method according to claim 1, characterized in that, The method further includes: The results of the target pulmonary hypertension detection are displayed. In addition to the pulmonary hypertension category identification results, the target pulmonary hypertension detection results also include the tricuspid regurgitation velocity to be detected by ultrasound and the corresponding lesion image features. Specifically, the lesion image features involve image features of pulmonary artery dilation, lung parenchymal changes, cardiac morphological changes, or chronic thromboembolism.

8. A pulmonary hypertension detection device integrating imaging features and tricuspid regurgitation velocity, characterized in that, The device includes: The acquisition unit is used to acquire the computed tomography image of the lungs to be detected and the tricuspid regurgitation velocity of the ultrasound corresponding to the current pulmonary hypertension detection task. The detection unit is used to input the lung computed tomography image to be detected and the tricuspid regurgitation velocity on ultrasound to be detected into a pre-selected pulmonary hypertension detection model. The pulmonary hypertension detection model includes a tissue site detection model and a pulmonary hypertension detection model. The tissue site detection model is used to classify and identify target tissue sites related to pulmonary hypertension based on the lung computed tomography image input into the model. The pulmonary hypertension detection model is used to combine the classification and identification results output by the tissue site detection model with the tricuspid regurgitation velocity on ultrasound input into the model to perform pulmonary hypertension detection processing. The resulting pulmonary hypertension detection processing result is used as the model output. The extraction unit is used to extract the target pulmonary hypertension detection results output by the pulmonary hypertension detection model.

9. A processing device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, and the processor executes the method as described in any one of claims 1 to 7 when it invokes the computer program in the memory.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a plurality of instructions adapted for loading by a processor to perform the method of any one of claims 1 to 7.

Images