Medical data visual propagation method and system based on 3D-AI technology

Abstract

The invention discloses a medical data visual propagation method and system based on a 3D-AI technology. The method comprises the following steps: firstly, acquiring multi-dimensional medical data from medical imaging equipment, reducing the dimension of the multi-dimensional medical data into a one-dimensional data sequence through a one-dimensional reconstruction algorithm, removing redundant information and retaining key features; thirdly, the one-dimensional data sequence is input into a deep learning model for training, and a three-dimensional model reflecting the medical data space structure is generated; and then, performing intelligent analysis on the three-dimensional model by using a pre-trained AI algorithm, extracting a feature mode of the lesion area, and generating target medical information. And finally, spreading the three-dimensional model and the analysis result to doctors, patients and other related personnel in a visual form through a network spreading technology. According to the method, the understandability of the medical data is improved through the deep learning technology, the diagnosis efficiency and accuracy are improved, the propagation efficiency of the medical data is optimized, and the method is suitable for processing and analyzing various medical image data.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data visualization, and particularly to a medical data visualization and dissemination method and system based on 3D-AI technology. Background Art

[0002] In the field of medical data visualization and dissemination, the existing technologies mainly rely on traditional two-dimensional image display and simple three-dimensional reconstruction technologies.

[0003] The two-dimensional image display method directly presents the two-dimensional slice images obtained by medical imaging devices (such as CT, MRI) to doctors and patients. However, this method lacks intuitive spatial information, making it difficult for doctors and patients to quickly and accurately grasp the overall condition of the disease. Especially for complex anatomical structures and lesions, it is difficult to understand. In addition, although the simple three-dimensional reconstruction technology can convert two-dimensional slice data into a three-dimensional model, it only constructs the model based on geometric information and pixel values, lacking in-depth exploration and analysis of the inherent characteristics of medical data, unable to use artificial intelligence algorithms to perform intelligent analysis on medical data, and difficult to discover early signs of diseases and the development trend of lesions.

[0004] In addition, the existing technologies have problems of low efficiency in medical data dissemination. They mainly rely on hospital internal systems or simple data sharing methods, lacking efficient visualization and dissemination means, resulting in restricted dissemination of medical information between different medical institutions, between doctors and patients. Therefore, the existing technologies have obvious deficiencies in data processing, model construction, data analysis, and dissemination efficiency, and cannot meet the requirements of modern medicine for efficient, accurate, and intelligent medical data visualization and dissemination. Summary of the Invention

[0005] Based on this, the embodiments of the present application provide a medical data visualization and dissemination method and system based on 3D-AI technology, which convert complex medical data into intuitive three-dimensional models and realize intelligent analysis and dissemination of data through AI algorithms.

[0006] In the first aspect, a medical data visualization and dissemination method based on 3D-AI technology is provided. The method includes:

[0007] Obtain multi-dimensional medical data collected by a medical imaging device, and perform dimensionality reduction processing on the multi-dimensional medical data through a one-dimensional reconstruction algorithm to convert it into a one-dimensional data sequence;

[0008] Input the one-dimensional data sequence and the corresponding annotation results into an artificial intelligence model based on deep learning for training to obtain a trained three-dimensional model; wherein, the three-dimensional model is used to generate a three-dimensional model reflecting the spatial structure of medical data according to the input one-dimensional data sequence;

[0009] Intelligently analyze the generated 3D model using a pre-trained AI algorithm to obtain target medical information;

[0010] Disseminate the 3D model and the analyzed target medical information to the target users in a visual form.

[0011] Optionally, the obtaining of multi-dimensional medical data collected by the medical imaging device includes:

[0012] Obtain two-dimensional slice image data from the medical imaging device;

[0013] Preprocess the two-dimensional slice image data, including removing noise, normalizing, and data augmentation;

[0014] Integrate the preprocessed two-dimensional slice image data into multi-dimensional medical data for subsequent one-dimensional reconstruction processing.

[0015] Optionally, the one-dimensional reconstruction algorithm uses a method based on Fourier transform. By performing Fourier transform on the multi-dimensional medical data, frequency features related to the spatial structure of the medical data are extracted, and the frequency features are transformed into a one-dimensional data sequence.

[0016] Optionally, the artificial intelligence model based on deep learning includes a convolutional neural network, and the convolutional neural network model is trained through the following steps:

[0017] Input the one-dimensional data sequence and the corresponding annotation results into the input layer of the convolutional neural network model;

[0018] Extract features from the input one-dimensional data sequence through convolutional layers and pooling layers;

[0019] Integrate the extracted features in the fully connected layer to generate a 3D model reflecting the spatial structure of the medical data;

[0020] Use the backpropagation algorithm to adjust the model parameters according to the annotation results and optimize the performance of the CNN model.

[0021] Optionally, the intelligent analysis of the generated 3D model using a pre-trained AI algorithm includes:

[0022] Extract features from the 3D model. The extracted features include the spatial position, size, shape of the lesion area, and the relative positional relationship with the surrounding tissues;

[0023] Input the extracted features into a pre-trained deep learning model. The model analyzes the features, identifies the feature patterns of the lesion area, and generates target medical information.

[0024] Optionally, the step of disseminating the three-dimensional model and the analyzed target medical information in a visual form to the target user includes:

[0025] Converting the three-dimensional model and the target medical information into a visual data format;

[0026] Transmitting the information in the visual data format to the terminal device of the target user through network dissemination technology;

[0027] Displaying the three-dimensional model and the target medical information on the terminal device in the form of a web end, a mobile end, or an AR / VR device.

[0028] In a second aspect, a medical data visualization dissemination system based on 3D-AI technology is provided, and the system includes:

[0029] A dimensionality reduction module, configured to obtain multi-dimensional medical data collected by a medical imaging device, and perform dimensionality reduction processing on the multi-dimensional medical data through a one-dimensional reconstruction algorithm to convert it into a one-dimensional data sequence;

[0030] A training module, configured to input the one-dimensional data sequence and the corresponding annotation results into an artificial intelligence model based on deep learning for training to obtain a trained three-dimensional model; wherein, the three-dimensional model is used to generate a three-dimensional model reflecting the spatial structure of medical data according to the input one-dimensional data sequence;

[0031] An analysis module, configured to perform intelligent analysis on the generated three-dimensional model by using a pre-trained AI algorithm to obtain target medical information;

[0032] A dissemination module, configured to disseminate the three-dimensional model and the analyzed target medical information in a visual form to the target user.

[0033] In a third aspect, an electronic device is provided, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the medical data visualization dissemination method according to any one of the first aspects is implemented.

[0034] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the medical data visualization dissemination method according to any one of the first aspects is implemented.

[0035] In a fifth aspect, a computer program product is provided, including a computer program / instructions, and when the computer program / instructions are executed by a processor, the medical data visualization dissemination method according to any one of the first aspects is implemented.

[0036] The beneficial effects brought by the technical solutions provided in the embodiments of the present application at least include:

[0037] (1) It can effectively extract the key features of the original medical data, remove redundant information, and provide a more accurate data basis for subsequent processing. For example, when processing complex brain MRI data, the one-dimensional reconstruction algorithm can quickly extract the key features related to brain lesions without being interfered by a large amount of useless background information, improving the accuracy and efficiency of subsequent model construction and analysis.

[0038] (2) The model learned from a large amount of medical data can accurately construct a three-dimensional model reflecting the spatial structure and features of medical data. Taking the diagnosis of liver diseases as an example, the three-dimensional model constructed by the deep learning model can clearly present the location and size of liver lesions and their spatial relationships with surrounding blood vessels and tissues, greatly reducing the difficulty for doctors and patients to understand medical data and reducing the possibility of misdiagnosis and missed diagnosis.

[0039] (3) AI algorithms can deeply explore the potential information in medical data. For example, in the early diagnosis of cancer, they can automatically extract subtle but key features that are difficult to detect by traditional techniques, providing more valuable diagnostic basis for doctors, helping doctors judge the condition earlier and more accurately, and formulating more effective treatment plans.

[0040] (4) It realizes the rapid and accurate transmission of medical information between different medical institutions, between doctors and patients. For example, when a patient is transferred, the receiving hospital can quickly obtain the complete and visual medical data of the patient through the network without the cumbersome manual handover and interpretation process, facilitating applications such as telemedicine and improving the efficiency and quality of medical services. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only exemplary, and for those of ordinary skill in the art, without creative efforts, other implementation drawings can be obtained according to the provided drawings.

[0042] Figure 1 It is a flowchart of the steps of a medical data visualization and transmission method based on 3D-AI technology provided by an embodiment of the present application;

[0043] Figure 2 It is a block diagram of a medical data visualization and transmission system based on 3D-AI technology provided by an embodiment of the present application;

[0044] Figure 3 It is a schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] In order to make the objectives, technical solutions and advantages of this application more clear and understandable, the following further elaborates on this application in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not used to limit this application.

[0046] In the description of the present invention, the terms "include", "have" and any of their variations are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units that are clearly listed, but may also include other steps or units that are inherent to these processes, methods, products or devices although not clearly listed, or steps or units added based on further optimized solutions conceived from the present invention.

[0047] In the prior art, in terms of the visualization and dissemination of medical data, traditional two-dimensional image display and simple three-dimensional reconstruction techniques are mainly used.

[0048] (1) Two-dimensional image display: directly present the two-dimensional slice images obtained by medical imaging devices (such as CT, MRI) to doctors and patients. Doctors need to rely on experience to integrate multiple two-dimensional images in their minds to build an understanding of the patient's body structure and lesions.

[0049] (2) Simple three-dimensional reconstruction techniques: Use surface rendering or volume rendering algorithms to convert two-dimensional slice data into three-dimensional models. However, these algorithms usually only consider the geometric information of the data and simply construct models based on pixel values. For example, in surface rendering, by extracting the boundaries of objects in the image and connecting these boundaries to form a three-dimensional surface model; volume rendering is to process the entire data volume and calculate its display effect in the final image according to the attributes of each voxel.

[0050] The prior art has the following disadvantages:

[0051] Difficulty in information understanding: In the two-dimensional image display mode, it is difficult for doctors and patients to quickly and accurately grasp the overall condition from a large number of two-dimensional slices because two-dimensional images lack intuitive spatial information, and it is more difficult to understand complex anatomical structures and lesions. For example, when diagnosing lung diseases, doctors need to judge the location, size and shape of lesions by themselves among many lung CT two-dimensional slices, which is prone to misdiagnosis or missed diagnosis.

[0052] Limited analysis ability: Simple three-dimensional reconstruction techniques only build models based on geometric information, lacking in-depth exploration and analysis of the inherent characteristics of medical data. It is impossible to use AI algorithms to perform intelligent analysis on medical data, and it is difficult to discover the early signs of diseases and the development trends of lesions. For example, in the early diagnosis of cancer, traditional techniques cannot automatically extract those subtle but crucial features from medical data.

[0053] Low propagation efficiency: In the existing technology for the dissemination of medical data, it mainly relies on the hospital's internal system or simple data sharing methods, lacking efficient visualization and dissemination means. This has restricted the dissemination of medical information between different medical institutions, as well as between doctors and patients. For example, when a patient is transferred, the process for the receiving hospital to obtain and quickly understand the patient's complete medical data is rather cumbersome.

[0054] The present invention utilizes one-dimensional reconstruction and artificial intelligence technologies to transform complex medical data into intuitive three-dimensional models, and realizes intelligent analysis and dissemination of the data through AI algorithms.

[0055] First, a specific one-dimensional reconstruction algorithm is adopted to perform dimensionality reduction on high-dimensional medical data and transform it into a one-dimensional data sequence. This step extracts and transforms the features of the original medical data, removes redundant information, and retains key features.

[0056] Next, the one-dimensional data sequence is input into an artificial intelligence model based on deep learning for training. After learning a large amount of medical data, this model can accurately construct a three-dimensional model that reflects the spatial structure and features of the medical data according to the one-dimensional data.

[0057] Then, the trained AI algorithm is used to perform intelligent analysis on the three-dimensional model to mine potential medical information therein, such as early signs of diseases and development trends of lesions.

[0058] Finally, through network dissemination technology, the analyzed medical data and three-dimensional models are disseminated to doctors, patients, and other relevant personnel in a visualized form through network platforms, mobile applications, etc.

[0059] Please refer to Figure 1 , which shows a flowchart of a method for visualizing and disseminating medical data based on 3D-AI technology provided by an embodiment of the present application, and may include the following steps:

[0060] S1. Obtain multi-dimensional medical data collected by a medical imaging device, and perform dimensionality reduction on the multi-dimensional medical data through a one-dimensional reconstruction algorithm to transform it into a one-dimensional data sequence.

[0061] In this step, multi-dimensional medical data is obtained from a medical imaging device (such as CT, MRI, PET, etc.). These data usually exist in the form of two-dimensional slice images, and each slice contains anatomical and pathological information of different levels of the patient's body. After obtaining the data, the following preprocessing operations are performed:

[0062] Noise removal: Use a filtering algorithm (such as Gaussian filtering or median filtering) to remove noise in the image and improve the data quality.

[0063] Normalization: Normalize the pixel values of the image data to a unified range (such as 0 to 1) for subsequent processing.

[0064] Data augmentation: Augment the image through operations such as rotation, flipping, and scaling to increase data diversity and improve the generalization ability of the model.

[0065] Data integration: Integrate the preprocessed two-dimensional slice image data into multi-dimensional medical data to form a complete medical image dataset for subsequent one-dimensional reconstruction processing.

[0066] Among them, the three-dimensional model is used to generate a three-dimensional model reflecting the spatial structure of medical data according to the input one-dimensional data sequence.

[0067] Use a one-dimensional reconstruction algorithm based on Fourier transform to perform dimensionality reduction on multi-dimensional medical data. The specific process is as follows:

[0068] Divide the multi-dimensional medical data into multiple small blocks, each small block containing medical information of a local area. Perform Fourier transform on each data block to extract its frequency domain features. Fourier transform can convert the data from the spatial domain to the frequency domain, facilitating the extraction of key features. In the frequency domain, identify the frequency components related to the spatial structure of medical data, remove redundant information, and retain key features. Encode the extracted frequency features into a one-dimensional data sequence to form the dimensionality-reduced data. For example, when processing brain MRI data, the algorithm can extract key features related to lesions and reduce the interference of irrelevant background information.

[0069] In an alternative embodiment of the present application, in terms of the one-dimensional reconstruction algorithm, the currently adopted one is the one-dimensional reconstruction algorithm based on Fourier transform. An alternative can be the one-dimensional reconstruction algorithm based on wavelet transform, which may have better effects when dealing with local features and can also achieve dimensionality reduction of medical data and convert it into a one-dimensional data sequence.

[0070] S2. Input the one-dimensional data sequence and the corresponding annotation results into an artificial intelligence model based on deep learning for training to obtain a trained three-dimensional model.

[0071] Input the one-dimensional data sequence and its corresponding annotation results into a deep learning model based on a convolutional neural network (CNN) for training. The specific process is as follows:

[0072] Input the one-dimensional data sequence and the corresponding annotation results (such as lesion location, type, etc.) into the input layer of the CNN model.

[0073] Extract features from the input one-dimensional data sequence through multiple convolutional layers and pooling layers. The convolutional layer is used to extract local features, and the pooling layer is used to reduce the feature dimension and retain key information.

[0074] Integrate the extracted features in the fully connected layer to generate a three-dimensional model that reflects the spatial structure of medical data. The model learns the labeled data to optimize the network parameters to ensure that the generated three-dimensional model can accurately reflect the spatial characteristics of medical data.

[0075] Use the backpropagation algorithm to adjust the model parameters according to the annotation results and optimize the performance of the CNN model. Through multiple iterative trainings, ensure that the model can accurately reconstruct the three-dimensional model from the one-dimensional data sequence.

[0076] S3. Use a pre-trained AI algorithm to perform intelligent analysis on the generated three-dimensional model to obtain target medical information.

[0077] The specific process of using a pre-trained AI algorithm to perform intelligent analysis on the generated three-dimensional model is as follows:

[0078] Extract features from the generated three-dimensional model. The extracted features include the spatial location, size, shape of the lesion area, and the relative positional relationship with the surrounding tissues.

[0079] Input the extracted features into a pre-trained deep learning model. The model analyzes the features to identify the feature patterns of the lesion area. For example, identify features such as the boundary, texture, and density of the lesion.

[0080] Generate target medical information based on the identified feature patterns, such as the type, severity, and development trend of the lesion. The analysis results are output in a structured form to provide a basis for doctors to make a diagnosis.

[0081] In an alternative embodiment of the present application, in terms of the AI model, the currently adopted one is a deep learning model based on a convolutional neural network (CNN). An alternative solution could be a model based on the Transformer architecture, which has advantages in processing long sequence data and may achieve better results in the analysis and modeling of medical data.

[0082] S4. Propagate the three-dimensional model and the analyzed target medical information to the target user in a visual form.

[0083] Convert the three-dimensional model and the analysis results into a visual data format, such as HTML5, OpenGL, or VRML, etc., for display on different devices.

[0084] Transmit the visual data to the target user's terminal device (such as a Web terminal, a mobile terminal, or an AR / VR device) through network propagation technologies (such as HTTP, FTP, etc.).

[0085] Display the 3D model and analysis results in an intuitive visual form on the terminal device. For example, display them through a browser plugin on the Web side, through an application on the mobile side, and through an immersive environment in AR / VR devices. Users can view medical data more intuitively through interactive operations (such as rotation, zoom, annotation, etc.).

[0086] In an alternative embodiment of the present application, the process is as follows:

[0087] Input: Obtain multi-dimensional medical data generated by medical imaging devices (such as CT, MRI).

[0088] One-dimensional reconstruction: Adopt a one-dimensional reconstruction algorithm based on Fourier transform to reduce the high-dimensional medical data to a one-dimensional data sequence, removing redundant information and retaining key features. For example, when processing brain MRI data, the algorithm can extract key features related to lesions and reduce the interference of irrelevant background information.

[0089] AI modeling and analysis: Input the one-dimensional data sequence into a deep learning model based on convolutional neural network (CNN) for training to build a high-precision 3D model. For example, in the diagnosis of liver diseases, the model can accurately present the location of the lesion and its spatial relationship with surrounding tissues.

[0090] Visualization dissemination: Through network dissemination technology, disseminate the 3D model and analysis results in a visual form (such as Web side, mobile side, AR / VR device) to doctors, patients and relevant personnel. For example, when a patient is referred, the receiving hospital can quickly obtain visual medical data through the network platform.

[0091] Output: 3D model, intelligent analysis report, visualization interface.

[0092] In summary, the differences and improvement effects compared with the prior art include:

[0093] (1) Differences and improvements in data processing methods:

[0094] Differences: The prior art mainly processes two-dimensional slice data before 3D reconstruction, and simple 3D reconstruction only constructs a model based on geometric information and pixel values; in this application, the high-dimensional medical data is first reduced to a one-dimensional data sequence through a specific one-dimensional reconstruction algorithm, removing redundant information and retaining key features.

[0095] Improvement effects: This method can extract the core features of medical data more efficiently, provide a better data basis for subsequent model construction and analysis, improve the accuracy and efficiency of data processing, and overcome the problem of insufficient data feature mining in the prior art.

[0096] (2) Differences and improvements in model construction technology:

[0097] Differences: The models constructed by existing simple 3D reconstruction technologies lack deep learning of the inherent characteristics of medical data. In this application, a one-dimensional data sequence is input into an artificial intelligence model based on deep learning for training. Through learning a large amount of medical data, a 3D model that can reflect the spatial structure and characteristics of medical data is constructed.

[0098] Improvement effects: The 3D model constructed based on deep learning is more accurate, can intuitively display the spatial structure and characteristics of medical data, reduces the difficulty for doctors and patients to understand medical data, reduces misdiagnosis and missed diagnosis, and solves the problem of difficult information understanding in the prior art.

[0099] (3) Differences and improvements in data analysis capabilities:

[0100] Differences: Existing technologies cannot use AI algorithms to intelligently analyze medical data. This application uses the trained AI algorithm to intelligently analyze the 3D model and mine potential medical information, such as early signs of diseases, trends of lesion development, etc.

[0101] Improvement effects: The intelligent analysis of the AI algorithm provides more valuable diagnostic basis for doctors, makes up for the limitation of the analysis capabilities of existing technologies, and helps doctors make more accurate diagnoses and treatment plans.

[0102] (4) Differences and improvements in dissemination methods:

[0103] Differences: Existing technologies rely on hospital internal systems or simple data sharing methods in the dissemination of medical data and lack efficient visualization dissemination means. This application uses network dissemination technology to disseminate the analyzed medical data and 3D model in a visual form to relevant personnel through network platforms, mobile applications, etc.

[0104] Improvement effects: It realizes the rapid and accurate dissemination of medical information between different medical institutions, between doctors and patients, facilitates patient referral and telemedicine, and solves the problem of low dissemination efficiency in the prior art.

[0105] The technical effects are as follows:

[0106] (1) At the data processing level:

[0107] Advantages: Existing technologies mostly directly process medical data based on two-dimensional slices or simple 3D reconstruction only relies on geometric and pixel information. In this invention, high-dimensional medical data is first reduced to a one-dimensional data sequence through a specific one-dimensional reconstruction algorithm.

[0108] Effect: The algorithm can effectively extract the key features of raw medical data, remove redundant information, and provide a more accurate data basis for subsequent processing. For example, when processing complex brain MRI data, the one-dimensional reconstruction algorithm can quickly extract the key features related to brain lesions without being disturbed by a large amount of useless background information, which improves the accuracy and efficiency of subsequent model construction and analysis.

[0109] (2) Model construction level:

[0110] Advantages: The model constructed by existing simple three-dimensional reconstruction technology is insufficient to learn the intrinsic characteristics of medical data; the present invention inputs one-dimensional data sequences into an artificial intelligence model based on deep learning for training.

[0111] Effect: The model learned from a large amount of medical data can accurately construct a three-dimensional model that reflects the spatial structure and characteristics of medical data. Taking liver disease diagnosis as an example, the three-dimensional model constructed by the deep learning model can clearly present the location and size of liver lesions and their spatial relationship with surrounding blood vessels and tissues, greatly reducing the difficulty for doctors and patients to understand medical data and the possibility of misdiagnosis and missed diagnosis.

[0112] (3) Data analysis level:

[0113] Advantages: The existing technology lacks the ability to intelligently analyze medical data; the present invention uses a trained AI algorithm to perform intelligent analysis on the three-dimensional model.

[0114] Effect: AI algorithms can deeply mine potential information in medical data. For example, in the early diagnosis of cancer, they can automatically extract subtle but critical features that are difficult to detect with traditional technologies, providing doctors with more valuable diagnostic evidence, helping doctors to diagnose the disease earlier and more accurately, and develop more effective treatment plans.

[0115] (4) Communication method:

[0116] Advantages: The existing technology relies on the internal hospital system or simple data sharing, and the means of dissemination are limited; the present invention uses network dissemination technology to disseminate the analyzed medical data and three-dimensional models in a visualized form through network platforms, mobile applications, etc.

[0117] Effect: It realizes the rapid and accurate dissemination of medical information between different medical institutions and between doctors and patients. For example, when a patient is transferred, the receiving hospital can quickly obtain the patient's complete and visualized medical data through the network without the need for tedious manual handover and interpretation processes, which facilitates applications such as telemedicine and improves the efficiency and quality of medical services.

[0118] From the above, it can be seen that the innovative points of this application include:

[0119] Combination of one-dimensional reconstruction and artificial intelligence technology: Applying one-dimensional reconstruction technology to medical data processing and combining it with artificial intelligence technology to achieve efficient processing and intelligent analysis of medical data is the core innovation point of this invention and is of great significance.

[0120] Three-dimensional model construction based on deep learning: Using deep learning algorithms to learn one-dimensional data and construct a high-precision three-dimensional model to accurately reflect the spatial structure and characteristics of medical data, which is of secondary importance.

[0121] Intelligent analysis and visual dissemination: Implementing intelligent analysis of medical data through AI algorithms and disseminating the results in a visual form to enhance the value and dissemination efficiency of medical information, which is also relatively crucial.

[0122] Such as Figure 2 , the embodiment of this application also provides a block diagram of a medical data visual dissemination system based on 3D-AI technology. The system may include:

[0123] A dimensionality reduction module, configured to obtain multi-dimensional medical data collected by a medical imaging device and perform dimensionality reduction processing on the multi-dimensional medical data through a one-dimensional reconstruction algorithm to convert it into a one-dimensional data sequence;

[0124] A training module, configured to input the one-dimensional data sequence and corresponding annotation results into an artificial intelligence model based on deep learning for training to obtain a trained three-dimensional model; wherein, the three-dimensional model is used to generate a three-dimensional model reflecting the spatial structure of medical data according to the input one-dimensional data sequence;

[0125] An analysis module, configured to perform intelligent analysis on the generated three-dimensional model using a pre-trained AI algorithm to obtain target medical information;

[0126] A dissemination module, configured to disseminate the three-dimensional model and the analyzed target medical information in a visual form to target users.

[0127] For the specific limitations of the medical data visual dissemination system based on 3D-AI technology, reference can be made to the limitations of the medical data visual dissemination method based on 3D-AI technology in the above text, which will not be elaborated here. Each module in the above-mentioned medical data visual dissemination system based on 3D-AI technology can be implemented in whole or in part through software, hardware, and their combinations. The above-mentioned modules can be embedded in or independent of the processor in a computer device in the form of hardware, or stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0128] In one embodiment, an electronic device is provided. The electronic device may be a computer, and its internal structure diagram may be as Figure 3As shown. The electronic device includes a processor, a memory, and a network interface connected via a system bus. Among them, the processor of the device is used to provide computing and control capabilities. The memory of the device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to transmit data for medical data visualization based on 3D-AI technology. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, it implements a method for medical data visualization transmission based on 3D-AI technology.

[0129] Those skilled in the art can understand that the structure shown in Figure 3 is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different component layout.

[0130] In one embodiment, there is also provided a computer-readable storage medium having a computer program stored thereon, which relates to all or part of the processes in the method of the above embodiment.

[0131] In one embodiment, there is also provided a computer program product including a computer program / instructions, which relates to all or part of the processes in the method of the above embodiment.

[0132] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided in this application can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or an external cache memory. By way of illustration and not limitation, RAM is available in M forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Symchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0133] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0134] The above embodiments only represent several implementation manners of this application. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent application. It should be noted that for those of ordinary skill in the art, without departing from the concept of this application, several modifications and improvements can still be made, and these all belong to the protection scope of this application. Therefore, the protection scope of this application patent should be subject to the appended claims.

Claims

1. A medical data visualization and dissemination method based on 3D-AI technology, characterized in that, The method includes: Obtaining multi-dimensional medical data collected by a medical imaging device, and performing dimensionality reduction processing on the multi-dimensional medical data through a one-dimensional reconstruction algorithm to convert it into a one-dimensional data sequence; Inputting the one-dimensional data sequence and the corresponding annotation results into an artificial intelligence model based on deep learning for training to obtain a trained three-dimensional model; wherein, the three-dimensional model is used to generate a three-dimensional model reflecting the spatial structure of medical data according to the input one-dimensional data sequence; Intelligently analyzing the generated three-dimensional model by using a pre-trained AI algorithm to obtain target medical information; Disseminating the three-dimensional model and the analyzed target medical information to the target user in a visual form.

2. The medical data visualization and dissemination method according to claim 1, wherein The obtaining of multi-dimensional medical data collected by a medical imaging device includes: Obtaining two-dimensional slice image data from a medical imaging device; Performing preprocessing on the two-dimensional slice image data, including noise removal, normalization processing, and data augmentation; Integrating the preprocessed two-dimensional slice image data into multi-dimensional medical data for subsequent one-dimensional reconstruction processing.

3. The medical data visualization and dissemination method according to claim 1, wherein The one-dimensional reconstruction algorithm adopts a method based on Fourier transform. By performing Fourier transform on the multi-dimensional medical data, frequency features related to the spatial structure of medical data are extracted, and the frequency features are converted into a one-dimensional data sequence.

4. The medical data visualization and dissemination method according to claim 1, characterized in that The artificial intelligence model based on deep learning includes a convolutional neural network, and the convolutional neural network model is trained through the following steps: Inputting the one-dimensional data sequence and the corresponding annotation results into the input layer of the convolutional neural network model; Performing feature extraction on the input one-dimensional data sequence through convolutional layers and pooling layers; Integrating the extracted features in the fully connected layer to generate a three-dimensional model reflecting the spatial structure of medical data; Using the backpropagation algorithm to adjust the model parameters according to the annotation results to optimize the performance of the CNN model.

5. The medical data visualization and dissemination method according to claim 1, characterized in that The intelligently analyzing the generated three-dimensional model by using a pre-trained AI algorithm includes: Performing feature extraction on the three-dimensional model, and the extracted features include the spatial position, size, shape of the lesion area, and the relative positional relationship with the surrounding tissues; Inputting the extracted features into a pre-trained deep learning model, and the model analyzes the features, identifies the feature patterns of the lesion area, and generates target medical information.

6. The medical data visualization and dissemination method according to claim 1, wherein The disseminating the three-dimensional model and the analyzed target medical information to the target user in a visual form includes: Converting the three-dimensional model and the target medical information into a visual data format; Transmitting the information in the visual data format to the terminal device of the target user through network dissemination technology; Displaying the three-dimensional model and the target medical information on the terminal device in the form of a Web terminal, a mobile terminal, or an AR / VR device.

7. A medical data visualization and dissemination system based on 3D-AI technology, characterized in that, The system includes: A dimensionality reduction module, configured to obtain multi-dimensional medical data collected by a medical imaging device, and perform dimensionality reduction processing on the multi-dimensional medical data through a one-dimensional reconstruction algorithm to convert it into a one-dimensional data sequence; A training module, configured to input the one-dimensional data sequence and the corresponding annotation results into an artificial intelligence model based on deep learning for training to obtain a trained three-dimensional model; wherein, the three-dimensional model is used to generate a three-dimensional model reflecting the spatial structure of medical data according to the input one-dimensional data sequence; An analysis module, configured to perform intelligent analysis on the generated three-dimensional model by using a pre-trained AI algorithm to obtain target medical information; A dissemination module, configured to disseminate the three-dimensional model and the analyzed target medical information to the target user in a visual form.

8. An electronic device, characterized in that, It includes a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that, A computer program is stored thereon, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images