A medical data aggregation system based on heterogeneous multiple data sources

Abstract

The present invention relates to the field of data processing technology, and specifically discloses a medical data aggregation system based on heterogeneous multi-data sources. Through a time dimension fusion unit and a space dimension fusion unit in a medical source fusion module, ultrasound contrast-enhanced gastrointestinal stromal tumor images and computer tomography gastrointestinal stromal tumor images acquired at different times are sequentially fused and fused in the same coordinate system, so that blood flow image information data provided by ultrasound contrast-enhanced imaging and anatomical structure image information data provided by computer tomography are combined to form a medical data aggregation system.

Description

Technical Field

[0001] The invention relates to the technical field of data processing, and specifically discloses a medical data aggregation system based on heterogeneous multiple data sources. Background Art

[0002] Gastrointestinal stromal tumors originate from gastrointestinal stromal cells and are common in the stomach and small intestine. Due to their heterogeneity and potential malignant tendency, it is important to accurately identify gastrointestinal stromal tumors through imaging. Ultrasound angiography, as a common imaging examination method, can provide clues for the initial discovery of gastrointestinal stromal tumors, but its penetration is limited and the imaging effect of lesions located deep in the gastrointestinal tract is poor. Computed tomography can provide high-resolution cross-sectional images and can clearly display the deep gastrointestinal structure, but its ability to evaluate the hemodynamics of gastrointestinal stromal tumors is poor, and it is difficult to fully reflect the blood supply characteristics of gastrointestinal stromal tumors. Existing identification of gastrointestinal stromal tumors can often only be carried out through a single imaging method, and the examination of gastrointestinal stromal tumors is not comprehensive. In order to solve the above problems, a technical solution is now provided. Summary of the invention

[0003] In order to overcome the above-mentioned defects of the prior art, the present invention provides a medical data aggregation system based on heterogeneous multiple data sources, which collects ultrasound contrast imaging gastrointestinal stromal tumor image data and computer tomography gastrointestinal stromal tumor image data through an endogenous acquisition module, and the medical source fusion module fuses heterogeneous data from different data sources. The medical research and analysis module identifies and segments the tumor area of ​​the gastrointestinal stromal tumor, and the visualization module displays the analysis results and influencing data to the doctor in an intuitive manner, so as to solve the problems raised in the above-mentioned background technology.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A medical data aggregation system based on heterogeneous multi-data sources includes an endogenous acquisition module, a medical source refining module, a medical source storage module, a medical source fusion module, a medical research identification module, a visualization module and a safety protection module, wherein the endogenous acquisition module acquires ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computer tomography gastrointestinal stromal tumor image data from an ultrasound device and a computer tomography device; the medical source refining module performs image preprocessing on the acquired ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computer tomography gastrointestinal stromal tumor image data; the medical source fusion module includes a time dimension fusion unit and a space dimension fusion unit, the time dimension fusion unit performs sequence fusion on ultrasound contrast-enhanced gastrointestinal stromal tumor images and computer tomography gastrointestinal stromal tumor images acquired during a patient examination cycle, constructs a time series minimum cumulative distance model, and measures the first time series. and the second time series The formula for the minimum cumulative distance model of time series is:

[0006] ,

[0007] Where: is the minimum cumulative distance value, The first time series and the second time series The optimal path to align between the first time series , The first time series The first element in The first time series The second element in The first time series The elements, the second time series , The second time series The first element in The second time series The second element in The second time series The elements, The first time series Mid-time point The index of The second time series Mid-time point The index of For the first time series Starting point to , and from the second time series Starting point to The minimum cumulative distance, The optimal path The length of the minimum cumulative distance between the two time series output , measure the first time series and the second time series If the minimum cumulative distance value is lower than the preset distance value, the pixel-level fusion technology is used to merge the first time series and the second time series Fusion; the medical research analysis module identifies and segments the tumor area of ​​gastrointestinal stromal tumors; the visualization module displays the analysis results and impact data to doctors in an intuitive way.

[0008] As a further solution of the present invention, the medical source fusion module includes a time dimension fusion unit and a space dimension fusion unit. The space dimension fusion unit fuses the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computed tomography gastrointestinal stromal tumor image in the same spatial coordinate system. The specific steps are as follows:

[0009] Step Y1, establishing a transformation model from the ultrasound contrast-enhanced gastrointestinal stromal tumor image coordinate system to the computed tomography gastrointestinal stromal tumor image coordinate system, the transformation model formula is:

[0010] ,

[0011] Where: is the point coordinate in ultrasound contrast-enhanced gastrointestinal stromal tumor image, is the coordinate of the point in the CT GIST image corresponding to the ultrasound contrast-enhanced GIST image, is a rotation matrix describing the rotation of the ultrasound contrast-enhanced gastrointestinal stromal tumor image relative to the computed tomography gastrointestinal stromal tumor image, is the translation vector, describing the translation of the ultrasound contrast-enhanced gastrointestinal stromal tumor image relative to the computed tomography gastrointestinal stromal tumor image;

[0012] Step Y2, establish an optimization model for the optimal transformation parameters, the formula of the model is:

[0013] ,

[0014] Where: is the optimal rotation matrix, is the optimal translation vector, A parameter value for measuring the similarity between an ultrasound contrast-enhanced gastrointestinal stromal tumor image and a computer tomography gastrointestinal stromal tumor image, when the parameter value is within a preset parameter value range, the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image are fused;

[0015] Step Y3, establishing a model of the fusion of ultrasound contrast-enhanced gastrointestinal stromal tumor images and computed tomography gastrointestinal stromal tumor images, the model formula is:

[0016] ,

[0017] Where: is the fused image, is the weighting factor, This is an ultrasound contrast-enhanced gastrointestinal stromal tumor image. Computed tomography image of a gastrointestinal stromal tumor.

[0018] As a further scheme of the present invention, the endogenous acquisition module is connected to the medical source refining module, the medical source refining module is connected to the medical source storage module, the medical source storage module is connected to the medical source fusion module, the time dimension unit and the space dimension fusion unit are respectively connected to the medical source fusion module, the medical source fusion module is connected to the medical research analysis module, the medical research identification module is connected to the visualization module, and the visualization module is connected to the security protection module.

[0019] As a further solution of the present invention, the endogenous acquisition module acquires ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computed tomography gastrointestinal stromal tumor image data from the ultrasound device and the computed tomography device. The endogenous acquisition module is connected to the ultrasound device and the computed tomography device through an interface. For the computed tomography device, a proprietary API is used to acquire image data. For the ultrasound device, a customized interface is used to acquire image data. The endogenous acquisition module automatically pulls gastrointestinal stromal tumor image data from the ultrasound device and the computed tomography device on demand and annotates patient information and examination results to ensure the accuracy of subsequent data management and retrieval. The gastrointestinal stromal tumor image data includes ultrasound contrast-enhanced gastrointestinal stromal tumor image data. and computed tomography gastrointestinal stromal tumor image data. Ultrasound contrast-enhanced gastrointestinal stromal tumor image data include dynamic image sequences and enhanced echo information. The dynamic image sequence appears in the form of a dynamic image sequence, showing the distribution and flow of microbubble contrast agents in the patient's gastrointestinal tract, helping to identify tumor blood flow characteristics. The enhanced echo information is that under the action of ultrasound, the contrast agent produces enhanced echoes; computed tomography gastrointestinal stromal tumor image data include high-resolution cross-sectional gastrointestinal stromal tumor images and volume parameters. The high-resolution cross-sectional gastrointestinal stromal tumor images are high-resolution two-dimensional cross-sectional images that can show the internal structure of the gastrointestinal tract in detail. Through multi-layer scanning, three-dimensional volume data can be generated for three-dimensional reconstruction and multi-view observation.

[0020] As a further solution of the present invention, the medical source refining module performs image preprocessing on the collected ultrasound contrast-enhanced gastrointestinal stromal tumor image data and the computer tomography gastrointestinal stromal tumor image data to adapt to subsequent data fusion and analysis, and uses a Gaussian kernel to convolve the image to eliminate the particle noise and speckle noise in the computer tomography gastrointestinal stromal tumor image and the ultrasound gastrointestinal stromal tumor image. The ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image are aligned through feature-based registration to make them consistent in space. The SIFT algorithm detects the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image. The key points in the gastrointestinal stromal tumor image are identified, and a scale-invariant feature vector is generated for each key point. The key point pairs in the two images are matched through feature description. The affine transformation between the images is estimated through the matched key point pairs. According to the estimated affine transformation matrix, the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image are resampled to align them in the same coordinate system. The global contrast of the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image is enhanced through histogram equalization. The second-order derivative of the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image is calculated using the Laplace operator to enhance the edge clarity in the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image.

[0021] As a further solution of the present invention, the medical source storage module stores multi-source heterogeneous data, wherein the multi-source heterogeneous data refers to ultrasound contrast-enhanced gastrointestinal stromal tumor image data, computer tomography gastrointestinal stromal tumor image data, patient information and examination results. The medical source storage module supports the management and access of large-scale data, and stores structured data and unstructured data. The structured data includes patient information and examination results, and the unstructured data includes ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computer tomography gastrointestinal stromal tumor image data. According to different requirements for data types, the system adopts a hybrid storage architecture, a relational database stores and manages structured data, and a NoSQL database stores and manages unstructured data.

[0022] As a further solution of the present invention, the medical research identification module identifies and segments the tumor area of ​​gastrointestinal stromal tumor. The medical research identification module uses the U-Net model to identify and segment the tumor area. The design of U-Net enables it to perform segmentation while retaining sufficient contextual information, which is particularly important for the identification and segmentation of gastrointestinal stromal tumors in medical images. Because the boundaries of gastrointestinal stromal tumors are relatively fuzzy and have different shapes, U-Net can accurately segment the gastrointestinal stromal tumor area at the pixel level. Based on the identification and segmentation results, the medical research identification module can generate quantitative analysis data of gastrointestinal stromal tumors, such as the size, shape, location of gastrointestinal stromal tumors and their relationship with surrounding tissues. When conducting medical research, the generated quantitative analysis data of gastrointestinal stromal tumors can provide researchers with rich quantitative analysis materials.

[0023] As a further solution of the present invention, a visualization module displays the analysis results and influencing data to the doctor in an intuitive manner. The module uses a volume rendering method to perform three-dimensional reconstruction on the fused ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computed tomography gastrointestinal stromal tumor image to display the spatial morphology and position of the gastrointestinal stromal tumor. The ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computed tomography gastrointestinal stromal tumor image corresponding to each time point on the timeline are then synchronously loaded into the playback window, and inter-frame interpolation is implemented to smooth the playback differences between different modalities. During the playback process, the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computed tomography gastrointestinal stromal tumor image are synchronously played through a unified time controller. The doctor can synchronously observe the condition of the gastrointestinal stromal tumor through the synchronous playback function of the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computed tomography gastrointestinal stromal tumor image.

[0024] The technical effects and advantages of the medical data aggregation system based on heterogeneous multiple data sources of the present invention are as follows:

[0025] The present invention collects ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computer tomography gastrointestinal stromal tumor image data through an endogenous acquisition module, pre-processes the collected images through a medical source refining module, stores multi-source heterogeneous data through a medical source storage module, fuses heterogeneous data from different data sources through a medical source fusion module to generate a unified patient view, identifies and segments tumor areas of gastrointestinal stromal tumors through a medical research analysis module, and displays analysis results and impact data to doctors in an intuitive manner through a visualization module. Combined with the blood flow image information data provided by ultrasound contrast-enhanced imaging and the anatomical structure image information data provided by computer tomography, a medical data aggregation system is formed, which helps to improve the accuracy of examining gastrointestinal stromal tumors. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 The present invention is a schematic diagram of the structure of a medical data aggregation system based on heterogeneous multiple data sources. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] Example 1

[0029] The embodiment of the present invention provides a medical data aggregation system based on heterogeneous multi-data sources, including an endogenous acquisition module, a medical source refining module, a medical source storage module, a medical source fusion module, a medical research identification module, a visualization module and a safety protection module. The medical source fusion module includes a time dimension fusion unit and a space dimension fusion unit. The endogenous acquisition module collects ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computer tomography gastrointestinal stromal tumor image data from an ultrasound device and a computer tomography device; the medical source refining module performs on the collected ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computer tomography gastrointestinal stromal tumor image data. Image preprocessing to adapt to subsequent data fusion and analysis; the medical resource storage module stores multi-source heterogeneous data, where multi-source heterogeneous data refers to ultrasound contrast-enhanced gastrointestinal stromal tumor image data, computed tomography gastrointestinal stromal tumor image data, patient information and examination results, supports large-scale data management and access, and stores structured and unstructured data; the medical source fusion module fuses heterogeneous data from different data sources to generate a unified patient view; the medical research analysis module identifies and segments tumor areas for gastrointestinal stromal tumors; the visualization module displays the analysis results and impact data to doctors in an intuitive way.

[0030] In this embodiment, the endogenous source acquisition module is connected to the medical source refining module, the medical source refining module is connected to the medical source storage module, the medical source storage module is connected to the medical source fusion module, the time dimension unit and the space dimension fusion unit are respectively connected to the medical source fusion module, the medical source fusion module is connected to the medical research analysis module, the medical research identification module is connected to the visualization module, and the visualization module is connected to the security protection module.

[0031] In this embodiment, the endogenous acquisition module acquires ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computed tomography gastrointestinal stromal tumor image data from the ultrasound device and the computed tomography device, including the following specific contents:

[0032] The endogenous acquisition module is connected to the ultrasound device and the computer tomography device through an interface. For the computer tomography device, a proprietary API is used to acquire image data. For the ultrasound device, a customized interface is used to acquire image data. The endogenous acquisition module is configured to automatically check whether the device has newly generated impact data after the patient examination is completed, and acquire gastrointestinal stromal tumor image data from the connected device;

[0033] The gastrointestinal stromal tumor image data includes ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computer tomography gastrointestinal stromal tumor image data. The ultrasound contrast-enhanced gastrointestinal stromal tumor image data includes dynamic image sequences and enhanced echo information. The dynamic image sequence appears in the form of a dynamic image sequence, showing the distribution and flow of microbubble contrast agents in the patient's gastrointestinal tract, helping to identify tumor blood flow characteristics, and the enhanced echo generates enhanced echo information under the action of ultrasound.

[0034] Computed tomography gastrointestinal stromal tumor image data include high-resolution cross-sectional gastrointestinal stromal tumor images and volume parameters. High-resolution cross-sectional gastrointestinal stromal tumor images are high-resolution two-dimensional cross-sectional images that can show the internal structure of the gastrointestinal tract in detail. Through multi-layer scanning, three-dimensional volume data can be generated for three-dimensional reconstruction and multi-view observation.

[0035] In this embodiment, the medical resource refining module performs image preprocessing on the collected ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computer tomography gastrointestinal stromal tumor image data to adapt to subsequent data fusion and analysis, including the following specific contents:

[0036] The collected ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computer tomography gastrointestinal stromal tumor image data were preprocessed to adapt to subsequent data fusion and analysis. The images were convolved with Gaussian kernels to eliminate particle noise and speckle noise in computer tomography gastrointestinal stromal tumor images and ultrasound gastrointestinal stromal tumor images. The ultrasound contrast-enhanced gastrointestinal stromal tumor images and computer tomography gastrointestinal stromal tumor images were aligned through feature-based registration to make them spatially consistent. The SIFT algorithm was used to detect key points in the ultrasound contrast-enhanced gastrointestinal stromal tumor images and computer tomography gastrointestinal stromal tumor images, and a scale-invariant feature vector was generated for each key point. The method matches key point pairs in two images, estimates the affine transformation between the images through the matched key point pairs, resamples the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image according to the estimated affine transformation matrix, aligns them in the same coordinate system, enhances the global contrast of the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image through histogram equalization, calculates the second-order derivative of the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image using the Laplace operator, and enhances the edge clarity in the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image.

[0037] In this embodiment, the medical resource storage module stores multi-source heterogeneous data, where the multi-source heterogeneous data refers to ultrasound contrast-enhanced gastrointestinal stromal tumor image data, computed tomography gastrointestinal stromal tumor image data, patient information and examination results, supports the management and access of large-scale data, and stores structured and unstructured data including the following specific contents:

[0038] Structured data includes patients’ personal information, medical history records, and examination results, which have clear formats and fields. Unstructured data includes ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computed tomography gastrointestinal stromal tumor image data, which exist in the form of images and videos.

[0039] To meet the storage requirements of different types of data, the medical resource storage module adopts a hybrid storage architecture. Structured data is stored and managed through a relational database, using its powerful transaction processing capabilities and complex query support to ensure data consistency and integrity. Unstructured data is stored and managed through a NoSQL database, which can flexibly process large-scale image data, support distributed storage and parallel processing, and improve the system's scalability and data access speed.

[0040] In this embodiment, the medical research identification module includes the following specific contents for identifying gastrointestinal stromal tumors and segmenting gastrointestinal stromal tumor regions:

[0041] The medical research recognition module uses the trained U-Net model to identify and segment the gastrointestinal stromal tumor area. The U-Net model consists of two parts: an encoder and a decoder. There is a jump connection between the encoder and the decoder. The design of U-Net enables it to retain sufficient contextual information while performing segmentation, which is particularly important for the recognition and segmentation of gastrointestinal stromal tumors in medical images. Because the boundaries of gastrointestinal stromal tumors are relatively fuzzy and their shapes vary, U-Net can accurately segment the gastrointestinal stromal tumor area at the pixel level. Based on the recognition and segmentation results, the medical research recognition module can generate quantitative analysis data of gastrointestinal stromal tumors, such as the size, shape, location of gastrointestinal stromal tumors and their relationship with surrounding tissues. When conducting medical research, the generated quantitative analysis data of gastrointestinal stromal tumors can provide researchers with rich quantitative analysis materials. The trained U-Net model weights and architecture are exported to TensorFlow format and deployed in the cloud server. The pixels of the input medical image data are normalized to eliminate the brightness differences under different scanning conditions. The processed medical image data are input into the U-Net model to obtain the identified and segmented gastrointestinal stromal tumor area.

[0042] In this embodiment, the medical source fusion module fuses heterogeneous data from different data sources to generate a unified patient view including the following specific contents:

[0043] The medical source fusion module has two units, namely the time dimension fusion unit and the space dimension fusion unit. The time dimension fusion unit sequentially fuses the ultrasound contrast-enhanced gastrointestinal stromal tumor images and computed tomography gastrointestinal stromal tumor images acquired at different times. The specific steps are as follows:

[0044] Step Z1, obtaining the first time series of ultrasound contrast-enhanced gastrointestinal stromal tumor images , The first time series The first time point in The first time series The second time point in The first time series The At the same time, a second time series of computed tomography gastrointestinal stromal tumor images was acquired , The second time series The first time point in The second time series The second time point in The second time series The time point;

[0045] Step Z2, first time series and the second time series The distance formula between two time points is: ,in, For time point and time point The distance between The first time series The time point, The second time series The time point;

[0046] Step Z3, the cumulative distance matrix formula is:

[0047] ,

[0048] Where: is the cumulative distance matrix, representing the distance from the first time series Starting point to The position and the time series from the second Starting point to The minimum cumulative distance of the location, The first time series The previous time point With the second time series Current time point The cumulative distance of The first time series Current time point With the second time series The previous time point The cumulative distance of The first time series The previous time point With the second time series The previous time point The cumulative distance of

[0049] Step Z4, construct the minimum cumulative distance model of the time series and measure the first time series and the second time series The formula for the minimum cumulative distance model of time series is:

[0050] ,

[0051] Where: is the minimum cumulative distance value, The optimal path for alignment between two time series, The first time series Mid-time point The index of The second time series Mid-time point The index of For the first time series Starting point to , and from the second time series Starting point to The minimum cumulative distance, The optimal path The length of the minimum cumulative distance between the two time series output , measure the first time series and the second time series If the minimum cumulative distance value is lower than the preset distance value, the pixel-level fusion technology is used to merge the first time series and the second time series Fusion;

[0052] The spatial dimension fusion unit fuses the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computed tomography gastrointestinal stromal tumor image in the same spatial coordinate system. The specific steps are as follows:

[0053] Step Y1, establishing a transformation model from the ultrasound contrast-enhanced gastrointestinal stromal tumor image coordinate system to the computed tomography gastrointestinal stromal tumor image coordinate system, the transformation model formula is:

[0054] ,

[0055] Where: is the point coordinate in ultrasound contrast-enhanced gastrointestinal stromal tumor image, is the coordinate of the point in the CT GIST image corresponding to the ultrasound contrast-enhanced GIST image, is a rotation matrix describing the rotation of the ultrasound contrast-enhanced gastrointestinal stromal tumor image relative to the computed tomography gastrointestinal stromal tumor image, is the translation vector, describing the translation of the ultrasound contrast-enhanced gastrointestinal stromal tumor image relative to the computed tomography gastrointestinal stromal tumor image;

[0056] Step Y2, establish an optimization model for the optimal transformation parameters, the formula of the model is:

[0057] ,

[0058] Where: is the optimal rotation matrix, is the optimal translation vector, A parameter value for measuring the similarity between an ultrasound contrast-enhanced gastrointestinal stromal tumor image and a computer tomography gastrointestinal stromal tumor image, when the parameter value is within a preset parameter value range, the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image are fused;

[0059] Step Y3, establishing a model of the fusion of ultrasound contrast-enhanced gastrointestinal stromal tumor images and computed tomography gastrointestinal stromal tumor images, the model formula is:

[0060] ,

[0061] Where: is the fused image, is the weighting factor, This is an ultrasound contrast-enhanced gastrointestinal stromal tumor image. Computed tomography image of a gastrointestinal stromal tumor.

[0062] In this embodiment, the visualization module displays the analysis results and impact data to the doctor in an intuitive manner, including the following specific contents:

[0063] The visualization module is a key part, responsible for presenting multimodal medical images to doctors in three-dimensional form. This module uses advanced volume rendering methods to perform high-precision three-dimensional reconstruction of fused ultrasound contrast-enhanced gastrointestinal stromal tumor images and computed tomography gastrointestinal stromal tumor images. This process can accurately present the spatial morphology, size and location of gastrointestinal stromal tumors, and highlight the differences between gastrointestinal stromal tumor tissues and surrounding tissues through different color and transparency settings;

[0064] The visualization module supports time-synchronized playback, which can display ultrasound contrast images and computed tomography gastrointestinal stromal tumor images collected at different time points in the same visualization interface. In the front-end interface, doctors can play, pause, and adjust the frame rate through the playback controller to flexibly control the playback rhythm of the image. The image data corresponding to each time point on the timeline will be synchronously loaded into the playback window. The system also uses inter-frame interpolation technology to smooth the time difference between different modalities to ensure the continuity and consistency of the playback process.

[0065] In addition, the visualization module integrates a unified time controller to ensure the precise synchronization of ultrasound contrast and computed tomography images of gastrointestinal stromal tumors. In actual operation, doctors can use this function to simultaneously observe and analyze ultrasound contrast and computed tomography images of gastrointestinal stromal tumors to obtain more comprehensive diagnostic information. The visualization module also provides a multi-viewing function, allowing doctors to rotate and scale the three-dimensional reconstructed images to view the morphology and position changes of gastrointestinal stromal tumors in detail from different angles.

[0066] The present invention collects image data of gastrointestinal stromal tumors from ultrasound equipment and computer tomography equipment through an endogenous acquisition module, pre-processes these image data through a medical source refining module for subsequent analysis and fusion, manages structured and unstructured data through a medical source storage module using a hybrid storage architecture, supports the management and access of large-scale data, and fuses heterogeneous data from different data sources in time and space dimensions through a medical source fusion module to generate a unified patient view, identifies and segments gastrointestinal stromal tumor areas through a medical source identification module, and generates quantitative analysis data, performs three-dimensional reconstruction through a visualization module, displays the fused image data, and supports time-synchronized playback function, so that doctors can intuitively observe the analysis results. The system adopts a hybrid storage architecture, can efficiently manage and access large-scale structured and unstructured data, and the three-dimensional reconstruction and time-synchronized playback functions enable doctors to intuitively observe the lesion condition. The system fuses heterogeneous data from multiple medical devices to generate a unified patient view.

[0067] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0068] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A medical data aggregation system based on heterogeneous multi-data sources, including an endogenous data collection module, a medical data refinement module, a medical data storage module, a medical data fusion module, a medical research identification module, a visualization module and a security protection module, characterized in that: The endogenous acquisition module collects ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computed tomography gastrointestinal stromal tumor image data from ultrasound equipment and computed tomography equipment; the medical source refining module performs image preprocessing on the collected ultrasound contrast-enhanced gastrointestinal stromal tumor image data and computed tomography gastrointestinal stromal tumor image data; the medical source fusion module includes a time dimension fusion unit and a space dimension fusion unit. The time dimension fusion unit performs sequence fusion on the ultrasound contrast-enhanced gastrointestinal stromal tumor images and computed tomography gastrointestinal stromal tumor images acquired during the patient's examination cycle, constructs a time series minimum cumulative distance model, and measures the first time series. and the second time series The formula for the minimum cumulative distance model of time series is: ， Where: is the minimum cumulative distance value, The first time series and the second time series The optimal path to align between the first time series , For time series The first element in The first time series The second element in The first time series The elements, the second time series , The second time series The first element in The second time series The second element in The second time series The elements, For time series Mid-time point The index of For time series Mid-time point The index of From the time series Starting point to , and from the time series Starting point to The minimum cumulative distance, The optimal path The length of the minimum cumulative distance between the two time series output , measure the first time series and the second time series If the minimum cumulative distance value is lower than the preset distance value, the pixel-level fusion technology is used to merge the first time series and the second time series Fusion; The medical source fusion module includes a time dimension fusion unit and a space dimension fusion unit. The space dimension fusion unit fuses the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computed tomography gastrointestinal stromal tumor image in the same spatial coordinate system. The specific steps are as follows: Step Y1, establishing a transformation model from the ultrasound contrast-enhanced gastrointestinal stromal tumor image coordinate system to the computed tomography gastrointestinal stromal tumor image coordinate system, the transformation model formula is: ， Where: is the point coordinate in ultrasound contrast-enhanced gastrointestinal stromal tumor image, is the coordinate of the point in the CT GIST image corresponding to the ultrasound contrast-enhanced GIST image, is a rotation matrix describing the rotation of the ultrasound contrast-enhanced gastrointestinal stromal tumor image relative to the computed tomography gastrointestinal stromal tumor image, is the translation vector, describing the translation of the ultrasound contrast-enhanced gastrointestinal stromal tumor image relative to the computed tomography gastrointestinal stromal tumor image; Step Y2, establish an optimization model for the optimal transformation parameters, the formula of the model is: , Where: is the optimal rotation matrix, is the optimal translation vector, A parameter value for measuring the similarity between an ultrasound contrast-enhanced gastrointestinal stromal tumor image and a computer tomography gastrointestinal stromal tumor image, when the parameter value is within a preset parameter value range, the ultrasound contrast-enhanced gastrointestinal stromal tumor image and the computer tomography gastrointestinal stromal tumor image are fused; Step Y3, establishing a model of the fusion of ultrasound contrast-enhanced gastrointestinal stromal tumor images and computed tomography gastrointestinal stromal tumor images, the model formula is: , Where: is the fused image, is the weighting factor, This is an ultrasound contrast-enhanced gastrointestinal stromal tumor image. Computed tomography image of a gastrointestinal stromal tumor.

2. A medical data aggregation system based on heterogeneous multiple data sources according to claim 1, characterized in that The endogenous source acquisition module is connected to the medical source refining module, the medical source refining module is connected to the medical source storage module, the medical source storage module is connected to the medical source fusion module, the time dimension unit and the space dimension fusion unit are respectively connected to the medical source fusion module, the medical source fusion module is connected to the medical research analysis module, the medical research identification module is connected to the visualization module, and the visualization module is connected to the security protection module.

3. The medical data aggregation system based on heterogeneous multiple data sources according to claim 1 is characterized in that: The medical resource storage module stores multi-source heterogeneous data, where the multi-source heterogeneous data refers to ultrasound contrast-enhanced gastrointestinal stromal tumor image data, computed tomography gastrointestinal stromal tumor image data, patient information and examination results. It supports the management and access of large-scale data and stores structured and unstructured data. Structured data refers to patient information and examination results, and unstructured data refers to ultrasound contrast-enhanced and computed tomography gastrointestinal stromal tumor image data. According to the different requirements of data types, a hybrid storage architecture is adopted, using a relational database to store and manage structured data and a NoSQL database to store and manage unstructured data.

Images