Cloud collaborative medical data management system and use method thereof

Abstract

The invention relates to the technical field of management, and particularly discloses a cloud collaborative medical data management system and a use method thereof, and the system comprises an image data collection module, a cloud storage module, a safety protection module and a collaborative management module. Prostate ultrasound image data are collected from personal equipment of medical institutions and doctors through the image data collection module and uploaded to the cloud storage module, the user authority management module provides user identity verification, authority management and behavior recording functions, and the safety protection module adopts an end-to-end encryption mode. And a gradient descent algorithm is used by the collaborative management module to obtain an optimal solution of a real-time collaborative operation queue model objective function, so that efficient cloud collaborative processing in the prostate ultrasonic image data management system is realized.

Description

Technical Field

[0001] The present invention relates to the field of management technologies, and more specifically, to a cloud collaborative medical data management system and a method for using the same. Background Art

[0002] Traditional prostate ultrasound image data management systems usually adopt the method of local storage and processing, which have limited storage space and difficult data sharing. There is a lack of effective technical means for cross-departmental and cross-institutional data collaboration, which easily leads to the phenomenon of information islands and affects the efficient utilization of prostate ultrasound image data. Although the existing technologies have begun to attempt to introduce cloud computing technologies to improve the efficiency of prostate ultrasound image data management, there are problems of low efficiency in the cloud collaborative processing in the prostate ultrasound image data management system.

[0003] In order to solve the above problems, a technical solution is provided now. Summary of the Invention

[0004] To overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a cloud collaborative medical data management system and a method for using the same. The prostate ultrasound image data is collected from medical institutions and doctors' personal devices by an image data acquisition module and uploaded to a cloud storage module. The user permission management module provides functions of user identity authentication, permission management, and behavior recording. The security protection module uses end-to-end encryption to ensure that the data is always in an encrypted state throughout the process from the sending end to the receiving end. The collaborative management uses the gradient descent algorithm to obtain the optimal solution of the objective function of the real-time collaborative operation queue model to solve the problems raised in the above background art.

[0005] To achieve the above object, the present invention provides the following technical solution: A cloud collaborative medical data management system includes an image data acquisition module, a cloud storage module, a data processing module, a user permission management module, a security protection module, and a collaborative management module. The image data acquisition module collects prostate ultrasound image data and uploads it to the cloud storage module. The data processing module includes a data analysis unit and an image processing function unit, supporting parallel processing of data. The user permission management module provides functions of user identity authentication, permission management, and behavior recording. The security protection module uses end-to-end encryption to ensure that the data is always in an encrypted state throughout the process from the sending end to the receiving end. The collaborative management module uses the gradient descent algorithm to obtain the optimal solution of the objective function of the real-time collaborative operation queue model, realizing high-efficiency cloud collaboration processing in the cloud collaborative medical data management system. The formula of the objective function of the real-time collaborative operation queue model is: , In the formula: The objective function of the real-time collaborative operation queue model is is the weight coefficient of the average response time, is the weight coefficient of the throughput, is the weight coefficient of the memory resource utilization rate, is the total number of data sharing and real-time collaborative operations, is the index variable of data sharing and real-time collaborative operations, is the average response time, is the weight related to the dependency relationship of data sharing and real-time collaborative operations, is the response time of the [th] data sharing and real-time collaborative operation, is the throughput, is the total time required to complete all data sharing and real-time collaborative operations, is the contribution factor of the [th] data sharing and real-time collaborative operation to the model throughput, is the memory resource utilization rate, is the total number of memory resource types, is the memory resource type, is the memory resource type total amount, is the specific demand of the [th] data sharing and real-time collaborative operation for the memory resource , is the collaborative operation queue; Using the gradient descent algorithm, the partial derivative of the objective function with respect to the average response time is , the partial derivative of the objective function with respect to the throughput is , the partial derivative of the objective function with respect to the memory resource utilization rate is , and the formula for updating variables according to the gradient descent method is:

[0006]

[0007] , where: is the learning rate, set to 0.01, the response time of the [th] specific data sharing and real-time collaborative operation, is the updated response time of the [th] data sharing and real-time collaborative operation, is the total time required to complete all data sharing and real-time collaborative operations, The total time required to complete all data sharing and real-time collaboration operations after the update, For the th data sharing and real-time collaboration operation, the specific demand for memory resources is For the th data sharing and real-time collaboration operation after the update, the specific demand for memory resources is. Repeat the above steps, continuously update the variables, and dynamically determine the stop of iteration by detecting the change rate of the objective function in each iteration. When the change rate of the objective function is less than for 10 consecutive iterations, stop the iteration, and the objective function reaches the optimal solution.

[0008] As a further solution of the present invention, the image data acquisition module is connected to the cloud storage module, the data processing module is connected to the cloud storage module, the user permission management module is connected to the data processing module, the security protection module is connected to the user permission management module, and the collaborative management module is connected to the security protection module. As a further solution of the present invention, the image data acquisition module collects prostate ultrasound image data from different medical institutions and doctors' personal devices and uploads it to the cloud, including: The image data acquisition module docks with the medical information systems of each institution, adopts the internationally common HL7 medical information exchange standard, and stably obtains prostate ultrasound image data from the servers of each medical institution; for doctors' personal devices, the image data acquisition module provides a dedicated application program, which supports Windows, IOS, and Android operating systems. Doctors upload the prostate ultrasound image data stored in their personal devices to the image data acquisition module by installing the dedicated application program; The collected prostate ultrasound image data has three formats: DICOM, JPEG, and PNG. The image data acquisition module uniformly converts the three different formats of prostate ultrasound image data. Using the OpenCV open-source image processing library, the prostate ultrasound images are converted from JPEG and PNG formats to DICOM format. It is set to automatically upload the prostate ultrasound image data converted to the unified DICOM format to the cloud at 2 am every day. During the upload process, the upload progress is displayed in real time, including the amount of data already uploaded and the remaining time information. Users can understand the progress of the upload through the progress bar. After the upload is completed, the image data acquisition module uses the hash algorithm to perform integrity verification on the uploaded data to ensure that the data is not lost or damaged during the transmission process.

[0009] As a further solution of the present invention, the cloud storage module provides the ability to store data and is composed of a distributed database unit and a cloud storage technology unit, including: The distributed database unit realizes high availability and fault tolerance of data by distributing prostate image data on multiple servers and nodes. The distributed database unit also includes data sharding, data replication, data consistency, availability, and partition tolerance. Data sharding divides data into multiple parts according to patient IDs, and each part is stored on different nodes. To improve data fault tolerance and availability, data is usually replicated to multiple nodes, and the distributed database needs to balance between consistency, availability, and partition tolerance; The cloud storage technology unit provides a storage solution with elasticity, scalability, and high availability. Cloud storage is combined with object storage services to manage large-scale unstructured data. Among them, unstructured data refers to prostate ultrasound images and patient information. The cloud storage technology unit stores data in a logically unified storage pool in the form of objects and uses redundancy technology to ensure high availability of data.

[0010] As a further solution of the present invention, the data processing module includes a data analysis unit and an image processing function unit, supporting parallel processing of large-scale data, including: The data analysis unit performs multi-dimensional quantitative analysis on prostate ultrasound images, reduces noise in prostate ultrasound images through median filtering, equalizes the gray histogram of prostate ultrasound images to enhance image contrast, uses the Canny edge detection algorithm to extract prostate boundary information in prostate ultrasound images, uses the U-Net segmentation algorithm based on deep learning to identify and segment nodules in prostate ultrasound images, and performs annotation. By performing three-dimensional reconstruction on multiple frames of prostate ultrasound images, the three-dimensional structure of the prostate is reconstructed to provide intuitive visualization information and establish a multi-frame ultrasound three-dimensional pixel reconstruction model. The formula for the multi-frame ultrasound three-dimensional pixel reconstruction model is: , In the formula: is the position in the horizontal direction of the prostate ultrasound image, is the position in the vertical direction of the prostate ultrasound image, is the depth coordinate in three-dimensional space, representing the position of the th frame of prostate ultrasound image in three-dimensional space, is a three-dimensional pixel value determined by the coordinates in three-dimensional space, is the pixel value of the th frame of prostate ultrasound image after image registration processing on the image plane. Among them, , is a two-dimensional affine transformation matrix. is the pixel value in the 2D prostate ultrasound image of the nth frame; The image processing functional unit uses a multi-core processor to decompose the image processing tasks into multiple threads for parallel execution, improving the processing efficiency. Through stream processing technology, it performs real-time processing on the prostate ultrasound image data and provides the processing results in a timely manner. It uses a GPU for large-scale matrix operations to accelerate the execution of the image processing algorithm.

[0011] As a further solution of the present invention, the user privilege management module provides user identity authentication, privilege management, and behavior monitoring, including: User identity authentication ensures the uniqueness of each user identity and the security of operations. The user privilege management module first allows users to register by providing basic information (such as username, password, email address, etc.) through the user registration mechanism, and at the same time supports the OAuth third-party authentication method to enhance the convenience and security of registration. The registration process includes preliminary verification of user information, such as sending verification emails or SMS verification codes, to ensure the authenticity of the user identity. When the user logs in, the system uses a hash algorithm to encrypt and store the user password, and further strengthens the security through two-factor authentication. After the user enters the password, they need to obtain and enter a one-time verification code through their mobile phone or email to complete the login. The password management function allows users to change their passwords at any time and provides the function of retrieving passwords. The retrieval process is verified through the email or mobile phone number provided during user registration. In addition, the user privilege management module limits the number of failed login attempts. After multiple consecutive failures, the account will be temporarily locked to prevent brute-force attack, thus comprehensively ensuring the security of user identity and the overall security of the system; Permission management assigns specific roles to each user to manage permissions. Different roles correspond to different operation permission sets. The user permission management module predefines multiple roles, such as ordinary users, doctors, and administrators, covering various operation permissions from basic data viewing to advanced data management. In addition, it also allows administrators to create custom roles and permissions according to specific needs and provides flexible permission allocation methods. It can be managed in batches by user groups or exclusive permissions can be set for individual users. The allocation and change of permissions follow an automatic inheritance mechanism. When the role permissions are modified, the permissions of all associated users will be updated synchronously to ensure the timeliness and consistency of management. To prevent permission abuse, the user permission management module implements a strict permission isolation mechanism. Low-permission users cannot access high-permission data beyond their authority. At the same time, the user permission management module includes detailed operation log records and real-time abnormal behavior monitoring functions. Each permission change and operation behavior will be recorded and audited. Administrators can track and analyze user behaviors at any time. Combining data encryption, permission isolation, and multi-level security protection, the user permission management module ensures the security and efficiency of permission management; Behavior monitoring records every operation of users in detail. These records include the operation time, user identity information, and specific operation objects.

[0012] As a further solution of the present invention, the security protection module uses end-to-end encryption to ensure that the data is always in an encrypted state throughout the process from the sending end to the receiving end, including: The prostate ultrasound image data is encrypted using an AES algorithm at the sending end. This encryption process converts the original plaintext data into a seemingly random and meaningless ciphertext form. The encryption key is only known to the sender and the receiver and is shared and negotiated in a secure manner. When the encrypted data is transmitted over the network and passes through routers and servers, since these intermediate nodes do not have the decryption key, they can only process the encrypted ciphertext data and cannot understand its true meaning. After receiving the encrypted data, the receiving end decrypts the ciphertext using the pre-shared or negotiated key to restore the original plaintext data. Throughout the process, the encryption and decryption operations of the data are only performed at the sending end and the receiving end, and no intermediate link can obtain the plaintext content of the data.

[0013] As a further solution of the present invention, a method for using cloud collaborative medical data management specifically includes the following steps: Step S1, collect prostate ultrasound image data from different medical institutions and doctors' personal devices through the image data acquisition module and upload it to the cloud storage module; Step S2, provide user identity authentication, permission management, and behavior recording functions through the user permission management module; Step S3: The collaborative management module establishes a real-time collaborative operation queue model, stores the prostate ultrasound image data in the cloud server, enables multiple users to access the prostate ultrasound image data in the cloud, and obtains the optimal solution of the objective function of the real-time collaborative operation queue model by using the gradient descent algorithm.

[0014] Technical effects and advantages of a cloud collaborative medical data management system and its usage method according to the present invention: The present invention collects prostate ultrasound image data from different medical institutions and doctors' personal devices through the image data acquisition module and uploads it to the cloud storage module. The user permission management module provides user identity authentication, permission management, and behavior recording functions. The security protection module uses end-to-end encryption to ensure that the data is always encrypted throughout the process from the sender to the receiver. The collaborative management module optimizes the objective function of the established real-time collaborative operation queue model, realizes efficient cloud collaborative processing while ensuring data security, and enables multiple users to perform concurrent operations on the shared data. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a schematic structural diagram of a cloud collaborative medical data management system according to the present invention; Figure 2 is a schematic flowchart of a usage method of a cloud collaborative medical data management according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0017] Embodiment 1 A cloud collaborative medical data management system of the present invention includes an image data acquisition module, a cloud storage module, a data processing module, a user permission management module, a security protection module, and a collaborative management module. The image data acquisition module acquires prostate ultrasound image data from different medical institutions and doctors' personal devices and uploads it to the cloud storage module; the data processing module includes a data analysis unit and an image processing function unit, supporting parallel processing of large-scale data; the user permission management module provides user identity authentication, permission management, and behavior recording functions; the security protection module uses an end-to-end encryption method to ensure that the data is always in an encrypted state throughout the process from the sending end to the receiving end; the collaborative management module optimizes the objective function of the established real-time collaborative operation queue model for multiple users to perform concurrent operations on shared data. The image data acquisition module is connected to the cloud storage module, the data processing module is connected to the cloud storage module, the user permission management module is connected to the data processing module, the security protection module is connected to the user permission management module, and the collaborative management module is connected to the security protection module. Further, the image data acquisition module acquires prostate ultrasound image data from medical institutions and doctors' personal devices and uploads it to the cloud, including: the image data acquisition module docks with the medical information systems of each institution, adopts the internationally common HL7 medical information exchange standard, and stably obtains prostate ultrasound image data from the servers of each medical institution; for doctors' personal devices, the image data acquisition module provides a dedicated application program that supports Windows, IOS, and Android operating systems. Doctors upload the prostate ultrasound image data stored in their personal devices to the image data acquisition module by installing the dedicated application program. The acquired prostate ultrasound image data has three formats: DICOM, JPEG, and PNG. The image data acquisition module uniformly converts the three different formats of prostate ultrasound image data, uses the OpenCV open-source image processing library to convert the prostate ultrasound images from JPEG and PNG formats to DICOM format, and sets the automatic upload of the prostate ultrasound image data converted to the unified DICOM format to the cloud at 2 am every day. During the upload process, the upload progress is displayed in real time, including the amount of data already uploaded and the remaining time information. Users can understand the upload progress through the progress bar. After the upload is completed, the image data acquisition module uses the hash algorithm to perform integrity verification on the uploaded data to ensure that the data is not lost or damaged during transmission.

[0018] Further, the cloud storage module provides the ability to store data and is composed of a distributed database unit and a cloud storage technology unit, including: The distributed database unit realizes high availability and fault tolerance of data by distributing prostate imaging data across multiple servers and nodes. The distributed database unit also includes data sharding, data replication, data consistency, availability, and partition tolerance. Data sharding divides data into multiple parts according to patient IDs, and each part is stored on different nodes. To improve data fault tolerance and availability, data is usually replicated to multiple nodes. The distributed database needs to balance between consistency, availability, and partition tolerance; The cloud storage technology unit provides a storage solution with elasticity, scalability, and high availability. Cloud storage combines with object storage services to manage large-scale unstructured data, where unstructured data refers to prostate ultrasound images and patient information. The cloud storage technology unit stores data in a logically unified storage pool in the form of objects and uses redundancy technology to ensure high availability of data.

[0019] Furthermore, the data processing module includes a data analysis unit and an image processing function unit, which support parallel processing of large-scale data, including: The data analysis unit performs multi-dimensional quantitative analysis on prostate ultrasound images. It reduces the noise of prostate ultrasound images through median filtering, equalizes the grayscale histogram of prostate ultrasound images to enhance image contrast, uses the Canny edge detection algorithm to extract the prostate boundary information in prostate ultrasound images, adopts the U-Net segmentation algorithm based on deep learning to identify and segment nodules in prostate ultrasound images and performs annotation, reconstructs the three-dimensional structure of the prostate by performing three-dimensional reconstruction on multiple frames of prostate ultrasound images, provides intuitive visualization information, and establishes a multi-frame ultrasound three-dimensional pixel reconstruction model. The formula for the multi-frame ultrasound three-dimensional pixel reconstruction model is: , In the formula: is the position in the horizontal direction of the prostate ultrasound image, is the position in the vertical direction of the prostate ultrasound image, is the depth coordinate in three-dimensional space, representing the position of the th frame of prostate ultrasound image in three-dimensional space, is a three-dimensional pixel value determined by the coordinates in three-dimensional space, is the pixel value of the th frame of prostate ultrasound image on the image plane after image registration processing, where, , is the pixel value of the th frame of prostate ultrasound image on the image plane, is a two-dimensional affine transformation matrix, is the Pixel values in the frame two-dimensional prostate ultrasound image; The image processing functional unit uses a multi-core processor to decompose the image processing tasks into multiple threads for parallel execution, improving the processing efficiency. Through stream processing technology, it performs real-time processing on the prostate ultrasound image data, provides the processing results in a timely manner, and uses the GPU for large-scale matrix operations to accelerate the execution of the image processing algorithm.

[0020] In this embodiment, when performing large-scale matrix operations using the GPU, convolution operations can also be performed using the GPU. A 3x3 convolution kernel is designed and uploaded from the host to the GPU. To enable the GPU to access quickly during convolution calculations, a convolution operation is started on the GPU. All threads work in parallel, and each thread independently calculates a part of the image. This process greatly accelerates the calculation. After the convolution operation is completed, the processing result is copied back from the GPU memory to the host memory.

[0021] Furthermore, the user permission management module provides user identity authentication, permission management, and behavior monitoring, including: User identity authentication ensures the uniqueness of each user identity and the security of operations. The user permission management module first allows users to register by providing basic information (such as username, password, email address, etc.) through the user registration mechanism, and at the same time supports the OAuth third-party authentication method to enhance the convenience and security of registration. The registration process includes preliminary verification of user information, such as sending verification emails or SMS verification codes, to ensure the authenticity of the user identity. When the user logs in, the system encrypts and stores the user password using a hash algorithm and further strengthens security through two-factor authentication. The user needs to obtain and enter a one-time verification code via mobile phone or email after entering the password to complete the login. The password management function allows users to change the password at any time and provides the function of retrieving the password. The retrieval process is verified through the email or mobile phone number provided during user registration. In addition, the user permission management module limits the number of failed login attempts. After multiple consecutive failures, the account will be temporarily locked to prevent brute-force cracking attacks, thus comprehensively ensuring the security of user identities and the overall security of the system; Permission management assigns specific roles to each user to manage permissions. Different roles correspond to different sets of operation permissions. The user permission management module predefines multiple roles, such as ordinary users, doctors, and administrators, covering various operation permissions from basic data viewing to advanced data management. In addition, it allows administrators to create custom roles and permissions according to specific needs and provides flexible permission allocation methods. Permissions can be managed in batches by user groups or exclusive permissions can be set for individual users. The allocation and change of permissions follow an automatic inheritance mechanism. When role permissions are modified, the permissions of all associated users will be updated synchronously to ensure the timeliness and consistency of management. To prevent permission abuse, the user permission management module implements a strict permission isolation mechanism. Low-privilege users cannot access high-privilege data without authorization. At the same time, the user permission management module includes detailed operation log records and real-time abnormal behavior monitoring functions. Each permission change and operation behavior will be recorded and audited. Administrators can track and analyze user behavior at any time. Combining data encryption, permission isolation, and multi-level security protection, the user permission management module ensures the security and efficiency of permission management; Behavior monitoring records every operation of users in detail. These records include the operation time, user identity information, and specific operation objects. Rules and thresholds are set to detect abnormal behaviors. If the number of consecutive login failures of the same user exceeds 10 times within 1 minute, it is regarded as an abnormal behavior. It is stipulated that if the number of edits to the same data file exceeds 20 times within 1 hour, it is regarded as an abnormal behavior. The normal working hours are set from 9:00 to 20:00. More than 3 operations outside the normal working hours are regarded as abnormal behaviors. The data volume downloaded by a single user within 1 day is limited to 10GB. A download volume exceeding 10GB is regarded as an abnormal behavior. When an abnormal behavior is detected, the user permission management module will immediately issue an alarm and notify the administrator, enabling the administrator to understand the operation dynamics of users in a timely manner; Behavior monitoring also regularly analyzes the behavior data of users to generate behavior reports. These reports help administrators understand users' usage habits, operation patterns, and potential risks. By deeply analyzing the behavior data, loopholes and deficiencies in the permission settings and operation processes of the user permission management module are discovered, providing a basis for further optimization and improvement.

[0022] Furthermore, the security protection module adopts an end-to-end encryption method to ensure that the data is always in an encrypted state throughout the process from the sending end to the receiving end, including: The prostate ultrasound image data is encrypted using an AES algorithm at the sending end. This encryption process converts the original plaintext data into a seemingly random and meaningless ciphertext form. The encryption key is known only to the sender and the receiver and is shared and negotiated in a secure manner. When the encrypted data is transmitted over the network through routers and servers, since these intermediate nodes do not have the decryption key, they can only process the encrypted ciphertext data and cannot understand its true meaning. After receiving the encrypted data, the receiving end decrypts the ciphertext using the pre-shared or negotiated key to restore the original plaintext data. Throughout the process, the encryption and decryption operations of the data are only carried out at the sending end and the receiving end, and no intermediate link can obtain the plaintext content of the data.

[0023] Furthermore, the collaborative management module optimizes the objective function of the established real-time collaborative operation queue model by using the gradient descent algorithm, realizing efficient cloud collaborative processing in the cloud collaborative medical data management system for multiple users to perform concurrent operations on shared data, including: The collaborative management module establishes a real-time collaborative operation queue model and stores the prostate ultrasound image data in the cloud server, enabling multiple users to access the prostate ultrasound image data in the cloud. The formula for the real-time collaborative operation queue model is: , In the formula: is the real-time collaborative operation queue model, is the total number of data sharing and real-time collaborative operations, is the index variable of data sharing and real-time collaborative operations, is the average response time, is the weight related to the dependency relationship of data sharing and real-time collaborative operations, is the th response time of data sharing and real-time collaborative operations, is the throughput, is the total time required to complete all data sharing and real-time collaborative operations, is the th contribution factor of data sharing and real-time collaborative operations to the model throughput, is the memory resource utilization rate, is the total number of memory resource types, is the memory resource type, is the memory resource type total amount of, is the th specific demand of data sharing and real-time collaborative operations for the memory resource , is the collaborative operation queue; The objective function of the real-time collaborative operation queue model is as follows: , In the formula: is the weight coefficient of the average response time, is the weight coefficient of the throughput, is the weight coefficient of the memory resource utilization rate. To optimize the objective function, the gradient descent algorithm is used. The partial derivative of the objective function with respect to the average response time is , and the partial derivative of the objective function with respect to the throughput is , and the partial derivative of the objective function with respect to the memory resource utilization rate is . According to the formula for updating variables by the gradient descent method:

[0024]

[0025] , In the formula: is the learning rate, set to 0.01, The th response time of data sharing and real-time collaborative operation, is the updated th response time of data sharing and real-time collaborative operation, is the total time required to complete all data sharing and real-time collaborative operations, is the updated total time required to complete all data sharing and real-time collaborative operations, is the th specific demand for memory resources for data sharing and real-time collaborative operation, is the updated th specific demand for memory resources for data sharing and real-time collaborative operation. Repeat the above steps to continuously update the variables. By detecting the change rate of the objective function in each iteration, the iteration is dynamically determined to stop. When the change rate of the objective function is less than for 10 consecutive iterations, the iteration stops, and the objective function reaches the optimal solution.

[0026] The usage method of a cloud collaborative medical data management of the present invention specifically includes the following steps: Step S1, collect prostate ultrasound image data from different medical institutions and doctors' personal devices through the image data acquisition module and upload it to the cloud storage module; Step S2, provide user authentication, permission management, and behavior recording functions through the user permission management module; Step S3, the collaborative management module establishes a real-time collaborative operation queue model, stores the prostate ultrasound image data in the cloud server, enables multiple users to access the prostate ultrasound image data in the cloud, and obtains the optimal solution of the objective function of the real-time collaborative operation queue model by using the gradient descent algorithm.

[0027] As mentioned above, it is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed by this application can easily think of changes or substitutions, which should all be covered within the protection scope of this application. Therefore, the protection scope of this application shall be subject to the protection scope of the claims.

[0028] Finally: The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A cloud-based collaborative medical data management system, comprising an image data acquisition module, a cloud storage module, a data processing module, a user authority management module, a security protection module and a collaborative management module, characterized in that: The image data acquisition module collects prostate ultrasound image data and uploads it to the cloud storage module; the data processing module includes a data analysis unit and an image processing functional unit, which supports parallel data processing; the user authority management module provides user identity authentication, authority management, and behavior recording functions; the security protection module uses end-to-end encryption to ensure that the data is always encrypted throughout the entire process from the sender to the receiver; the collaborative management module uses the gradient descent algorithm to obtain the optimal solution of the objective function of the real-time collaborative operation queue model, and realizes efficient cloud collaborative processing in the cloud collaborative medical data management system. The formula of the objective function of the real-time collaborative operation queue model is: ， Where: The objective function of the real-time collaborative operation queue model is: is the weight coefficient of the average response time, is the throughput weight coefficient, is the weight coefficient of memory resource utilization, is the total number of data sharing and real-time collaborative operations, Index variables for data sharing and real-time collaborative operations. is the average response time, is the weight associated with the dependency between data sharing and real-time collaborative operations, For the data sharing and real-time collaborative operation response time, is the throughput, The total time required to complete all data sharing and real-time collaboration operations, For the The contribution factors of data sharing and real-time collaborative operations to model throughput, is the memory resource utilization, is the total number of memory resource types, is the memory resource type, Memory resource type The total amount, For the Data sharing and real-time collaborative operations on memory resources The specific demand A collaborative operation queue.

2. A cloud-based collaborative medical data management system according to claim 1, characterized in that ,Using the gradient descent algorithm, we can get the optimal solution of the objective function of the real-time collaborative operation queue model. The partial derivative with respect to the average response time is , the objective function The partial derivative with respect to throughput is , the objective function The partial derivative of memory resource utilization is , the formula for updating variables according to the gradient descent method is: ， Where: is the learning rate, set to 0.01, No. data sharing and real-time collaborative operation response time, For the updated data sharing and real-time collaborative operation response time, The total time required to complete all data sharing and real-time collaboration operations, The total time required to complete all data sharing and real-time collaboration operations after the update. For the Data sharing and real-time collaborative operations on memory resources The specific demand After the update Data sharing and real-time collaborative operations on memory resources The specific demand is repeated, and the variables are continuously updated. By detecting the objective function The rate of change in each iteration is dynamically determined to stop the iteration when the objective function The rate of change is less than for 10 consecutive iterations , stop the iteration, then the objective function Reach the optimal solution.

3. A cloud-based collaborative medical data management system according to claim 1, characterized in that , The image data acquisition module is connected to the cloud storage module, the data processing module is connected to the cloud storage module, the user authority management module is connected to the data processing module, the security protection module is connected to the user authority management module, and the collaborative management module is connected to the security protection module.

4. A cloud-based collaborative medical data management system according to claim 1, characterized in that: The data processing module includes a data analysis unit and an image processing functional unit, which supports parallel processing of large-scale data. The data analysis unit reconstructs the three-dimensional structure of the prostate by performing three-dimensional reconstruction on multiple frames of prostate ultrasound images, provides intuitive visualization information, and establishes a multi-frame ultrasound three-dimensional pixel reconstruction model. The formula of the multi-frame ultrasound three-dimensional pixel reconstruction model is: ， Where: is the horizontal position of the prostate ultrasound image, is the vertical position of the prostate ultrasound image, is the depth coordinate in three-dimensional space, representing the The position of the prostate ultrasound image frame in three-dimensional space, In three-dimensional space, the coordinates Determine a three-dimensional pixel value, After image registration processing The pixel value of the frame prostate ultrasound image on the image plane, where , For the The pixel value of the frame prostate ultrasound image on the image plane, is a two-dimensional affine transformation matrix, For the The pixel values ​​in the two-dimensional prostate ultrasound image of the frame; the image processing functional unit uses a multi-core processor to decompose the image processing task into multiple threads for parallel execution to improve processing efficiency. Through stream processing technology, real-time processing of prostate ultrasound image data is performed to provide processing results in a timely manner. The GPU is used for large-scale matrix operations to accelerate the execution of image processing algorithms.

5. A method for using cloud-based collaborative medical data management, used to implement a cloud-based collaborative medical data management system as described in any one of claims 1 to 4, characterized in that: Step S1, collecting prostate ultrasound image data through the image data collection module and uploading it to the cloud storage module; Step S2, providing user identity authentication, rights management and behavior recording functions through the user rights management module; Step S3, the collaborative management module establishes a real-time collaborative operation queue model, stores the prostate ultrasound image data in the cloud server, allows multiple users to access the prostate ultrasound image data in the cloud, and obtains the optimal solution of the objective function of the real-time collaborative operation queue model by using the gradient descent algorithm.

Images