Remote diagnosis and report generation method of multi-modal medical data

Abstract

The application discloses a remote diagnosis and report generation method for multi-modal medical data, belongs to the technical field of data security, and can effectively improve the data encryption degree and guarantee the safety of remote diagnosis by performing encryption processing on the multi-modal medical data after preprocessing through a deep symmetric encryption algorithm, obtaining the multi-modal medical data after encryption processing, and transmitting the multi-modal medical data after encryption processing to a cloud server, then performing identity verification on a diagnosis expert in response to an access request initiated by the diagnosis expert, and allowing the diagnosis expert to access the multi-modal medical data after encryption processing in the case that the identity verification is passed, finally obtaining a diagnosis result input by the diagnosis expert through man-machine online interaction, filling the diagnosis result into a preset diagnosis report template to generate a patient diagnosis report, and returning the patient diagnosis report to a device corresponding to the patient, so that remote diagnosis and report generation are realized.

Description

Technical Field

[0001] This application relates to the field of medical information technology, and more specifically, to a method for remote diagnosis and report generation of multimodal medical data. Background Technology

[0002] With the development of internet technology and medical informatization, telemedicine has become an important means to optimize the allocation of medical resources and improve the accessibility of medical services. In remote diagnosis scenarios, doctors need to access various types of patient medical data, such as CT and MRI images, pathological slides, gene sequencing data, electronic medical records, and other multimodal information. This data is highly sensitive and involves patient privacy; its secure transmission and storage are the core challenges of telemedicine systems.

[0003] Existing telemedicine systems typically employ standard encryption algorithms (such as AES (Advanced Encryption Standard) and RSA (asymmetric encryption)) to protect data. However, the key generation and management methods for these standard algorithms are relatively fixed, potentially posing security risks when facing increasingly sophisticated cyberattacks (such as quantum computing threats and side-channel attacks). Once the keys are cracked, patients' sensitive data will be at risk of being leaked, leading to serious consequences. Summary of the Invention

[0004] This application aims to provide a method for remote diagnosis and report generation of multimodal medical data, which addresses the technical problem of insufficient data security in the existing remote diagnosis process.

[0005] This application provides a method for remote diagnosis and report generation of multimodal medical data, including:

[0006] Collect multimodal medical data corresponding to the patient, and preprocess the multimodal medical data to obtain preprocessed multimodal medical data;

[0007] The preprocessed multimodal medical data is encrypted using a deep symmetric encryption algorithm to obtain encrypted multimodal medical data.

[0008] The encrypted multimodal medical data is transmitted to a cloud server, and the cloud server stores the encrypted multimodal medical data in chronological order.

[0009] In response to an authorized diagnostic expert's terminal initiating an access request to the cloud server, the diagnostic expert is authenticated, and if the authentication is successful, the diagnostic expert is allowed to access the encrypted multimodal medical data.

[0010] The diagnostic results are obtained from the diagnostic experts through online human-computer interaction, and the diagnostic results are filled into a preset diagnostic report template to generate a patient diagnostic report. The patient diagnostic report is then sent back to the patient's corresponding device.

[0011] In one possible implementation, a deep symmetric encryption algorithm is used to encrypt the preprocessed multimodal medical data to obtain encrypted multimodal medical data, including:

[0012] Generate the first-round key for the symmetric encryption key, pre-convert the first-round key, obtain the pre-converted first-round key, set the key round counter t to 1, and use the pre-converted first-round key as the target key for the t-th round;

[0013] Multiple candidate keys are generated, and the evaluation values ​​of the candidate keys are obtained. The best candidate key is obtained based on the evaluation values ​​of the candidate keys.

[0014] Based on the best candidate key, the candidate key is sequentially subjected to single-point surround exploration, combined random exploration, and hop-point random exploration to obtain the candidate key after exploration;

[0015] Based on the candidate keys obtained after exploration, an elite selection strategy is executed to obtain a selection key, and the best candidate key is updated based on the selection key to obtain the updated best candidate key.

[0016] Repeat the process of updating the best candidate key multiple times to determine the target key for round t+1. Then, determine whether the count value of the key round counter is equal to the number of rounds required for the symmetric encryption key. If so, use all target keys together as the symmetric encryption key. Otherwise, increment the count value of the key round counter by one and return to the step of generating multiple candidate keys.

[0017] The preprocessed multimodal medical data is encrypted using a symmetric encryption key to obtain encrypted multimodal medical data.

[0018] In one possible implementation, generating a first-round key for a symmetric encryption key and pre-converting the first-round key to obtain a pre-converted first-round key includes:

[0019] The first-round key for generating the symmetric encryption key is generated using the AES algorithm;

[0020] The first-round key is pre-converted into a decimal number to obtain the pre-converted first-round key.

[0021] In one possible implementation, multiple candidate keys are generated, evaluation values ​​of the candidate keys are obtained, and the best candidate key is obtained based on the evaluation values ​​of the candidate keys, including:

[0022] Generate multiple candidate keys;

[0023] Obtain the Euclidean distance variance between the candidate key and the target key in the previous t rounds to get the evaluation value of the candidate key;

[0024] Based on the evaluation values ​​of all candidate keys, the candidate key with the highest evaluation value is determined as the best candidate key.

[0025] In one possible implementation, the single-point surround exploration includes:

[0026]

[0027]

[0028]

[0029]

[0030] in, Indicates the first t Round i Candidate keys Indicates the first i Candidate keys after a single-point surround exploration Indicates the first t The average of all candidate keys in the round. Indicates the best candidate key. Indicates the first single-point surrounding exploration factor. This indicates the second single-point surrounding exploration factor. Indicates the first t Round i The inertial weights of each candidate key. Indicates the first learning factor. Indicates the second learning factor. This represents the exploration angle coefficient, and , This represents the first constant term between [5, 10]. Represents pi (π). Represents a uniformly distributed random number between [0,1]. Represents the exploration quantity coefficient, and , This represents the exploration cycle coefficient between [0.5, 2]. Represents the corresponding parameter vectors The maximum value in, Represents the corresponding parameter vectors The maximum value in; This represents the maximum value of the inertia weight. This represents the minimum inertial weight. Indicates the first t Round i The evaluation value of each candidate key. This represents the evaluation value of the best candidate key. This represents the evaluation value of the worst candidate key. express The corresponding evaluation value.

[0031] In one possible implementation, the combined random exploration includes

[0032]

[0033]

[0034]

[0035] in, Indicates the first m Candidate keys after a single-point surround exploration Indicates the first m Candidate keys after a combination of random explorations This represents the weighted key of the candidate keys after exploring all single-point perimeters. Indicates the best candidate key. Indicates the weighting coefficient. This indicates the total number of candidate keys.

[0036] In one possible implementation, the hop random exploration includes:

[0037]

[0038]

[0039] in, Indicates the first n Candidate keys after a combination of random explorations Indicates the first n Candidate keys after a random exploration of each hop. Indicates the best candidate key. This represents a random step size control factor between (0,1). Indicated by As the mean, with A Gaussian distributed random number with variance . Let represent the cosine function, and let represent pi.

[0040] In one possible implementation, based on the candidate keys obtained after exploration and the candidate keys themselves, an elite selection strategy is executed to obtain a selection key, and the best candidate key is updated based on the selection key to obtain the updated best candidate key, including:

[0041] Determine whether the evaluation value of the candidate key after exploration is greater than the evaluation value of its corresponding original candidate key. If so, determine the candidate key after exploration as the selection key; otherwise, determine the corresponding original candidate key as the selection key.

[0042] Determine the selection key with the largest evaluation value, and determine whether the evaluation value corresponding to the selection key with the largest evaluation value is greater than the evaluation value of the best candidate key. If so, the selection key with the largest evaluation value is used as the new best candidate key to obtain the updated best candidate key; otherwise, the original best candidate key is directly used as the updated best candidate key.

[0043] In one possible implementation, the preprocessed multimodal medical data is encrypted using a symmetric encryption key to obtain encrypted multimodal medical data, including:

[0044] Based on the symmetric encryption key, the preprocessed multimodal medical data is encrypted using the AES algorithm, and the symmetric encryption key is encrypted using the public key of the cloud server to obtain the encrypted multimodal medical data.

[0045] In one possible implementation, in response to an authorized diagnostic expert terminal initiating an access request to the cloud server, the diagnostic expert is authenticated, including:

[0046] In response to an authorized diagnostic expert terminal initiating an access request to the cloud server, the diagnostic expert undergoes at least one of facial recognition, iris recognition, and fingerprint recognition to obtain an authentication result; wherein the authentication result is either successful or unsuccessful.

[0047] Beneficial effects:

[0048] This application provides a method for remote diagnosis and report generation of multimodal medical data. It encrypts preprocessed multimodal medical data using a deep symmetric encryption algorithm, resulting in encrypted multimodal medical data, which is then transmitted to a cloud server. This effectively enhances data encryption and ensures the security of remote diagnosis. In response to access requests from diagnostic experts, the method authenticates the experts and, if successful, allows them access to the encrypted multimodal medical data. Finally, it obtains the diagnostic results input by the experts through online human-computer interaction, fills the results into a preset diagnostic report template to generate a patient diagnostic report, and sends the report back to the patient's corresponding device, thus realizing remote diagnosis and report generation. Attached Figure Description

[0049] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart of a method for remote diagnosis and report generation of multimodal medical data proposed in one embodiment of this application. Detailed Implementation

[0051] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0052] like Figure 1 As shown in the figure, this application provides a method for remote diagnosis and report generation of multimodal medical data, including:

[0053] S101. Collect multimodal medical data corresponding to the patient, and preprocess the multimodal medical data to obtain preprocessed multimodal medical data.

[0054] At medical institutions or patient sites, multimodal medical data is collected from patients using equipment such as CT (computed tomography), MRI (magnetic resonance imaging), ultrasound, and endoscopy. Preprocessing includes, but is not limited to: image denoising, format standardization (such as conversion to DICOM format), data anonymization (removal of direct identifiers such as names and ID numbers), and size normalization, to improve the efficiency of subsequent processing and protect basic privacy.

[0055] S102. The preprocessed multimodal medical data is encrypted using a deep symmetric encryption algorithm to obtain encrypted multimodal medical data.

[0056] Existing technologies often employ correlated generation methods to generate multiple rounds of keys during the symmetric encryption key generation process, resulting in a certain correlation between the round keys and making them vulnerable to cracking. Therefore, this application embodiment uses a deep symmetric encryption algorithm to encrypt the preprocessed multimodal medical data, reducing the correlation between the round keys and improving the security of the encrypted multimodal medical data.

[0057] S103. The encrypted multimodal medical data is transmitted to a cloud server, and the encrypted multimodal medical data is stored in chronological order through the cloud server.

[0058] The encrypted medical data and the encrypted symmetric key are transmitted together to the cloud server. The cloud server uses its own private key to decrypt the data to obtain the symmetric key, then uses this key to decrypt the medical data and stores it in an ordered manner according to patient ID and timestamp.

[0059] S104. In response to the authorized diagnostic expert terminal initiating an access request to the cloud server, the diagnostic expert is authenticated, and if the authentication is successful, the diagnostic expert is allowed to access the encrypted multimodal medical data.

[0060] Authorized diagnostic experts initiate access requests to the cloud via their terminals. The cloud server initiates an authentication process, requiring the expert to verify at least one biometric method, such as facial recognition, iris recognition, or fingerprint recognition. Only after successful verification can the expert gain access to the corresponding patient data.

[0061] Optionally, an artificial intelligence model can be deployed directly on a cloud server, and the model can automatically identify and process the multimodal medical data to obtain diagnostic results, thereby achieving automated assisted diagnosis.

[0062] S105. Obtain the diagnostic results input by the diagnostic expert through online human-computer interaction, fill the diagnostic results into a preset diagnostic report template to generate a patient diagnostic report, and send the patient diagnostic report back to the patient's corresponding device.

[0063] Experts can browse and analyze multimodal medical data online, and input diagnostic conclusions and treatment suggestions through an interactive interface. The system backend automatically fills this information into a preset diagnostic report template that conforms to medical standards, generating a complete PDF or HTML report.

[0064] This application provides a method for remote diagnosis and report generation of multimodal medical data. It encrypts preprocessed multimodal medical data using a deep symmetric encryption algorithm, resulting in encrypted multimodal medical data, which is then transmitted to a cloud server. This effectively enhances data encryption and ensures the security of remote diagnosis. In response to access requests from diagnostic experts, the method authenticates the experts and, if successful, allows them access to the encrypted multimodal medical data. Finally, it obtains the diagnostic results input by the experts through online human-computer interaction, fills the results into a preset diagnostic report template to generate a patient diagnostic report, and sends the report back to the patient's corresponding device, thus realizing remote diagnosis and report generation.

[0065] In one possible implementation, a deep symmetric encryption algorithm is used to encrypt the preprocessed multimodal medical data to obtain encrypted multimodal medical data, including:

[0066] Generate the first-round key for the symmetric encryption key, pre-convert the first-round key, obtain the pre-converted first-round key, set the key round counter t to 1, and use the pre-converted first-round key as the target key for the t-th round;

[0067] Multiple candidate keys are generated, and the evaluation values ​​of the candidate keys are obtained. The best candidate key is obtained based on the evaluation values ​​of the candidate keys.

[0068] Based on the best candidate key, the candidate key is sequentially subjected to single-point surround exploration, combined random exploration, and hop-point random exploration to obtain the candidate key after exploration;

[0069] Based on the candidate keys obtained after exploration, an elite selection strategy is executed to obtain a selection key, and the best candidate key is updated based on the selection key to obtain the updated best candidate key.

[0070] Repeatedly update the best candidate key multiple times (i.e., repeatedly perform single-point surround exploration, combined random exploration, and hop-point random exploration), determine the target key for round t+1, and determine whether the count value of the key round counter is equal to the required number of rounds for the symmetric encryption key. If so, use all target keys together as the symmetric encryption key; otherwise, increment the count value of the key round counter by one and return to the step of generating multiple candidate keys.

[0071] The preprocessed multimodal medical data is encrypted using a symmetric encryption key to obtain encrypted multimodal medical data.

[0072] The key is not generated all at once, but through a process of multiple iterations and continuous evolution. The target key in each round becomes the cornerstone of the next round of search, making the final key the cumulative result of the entire optimization process, with extremely high complexity and nonlinear characteristics.

[0073] This application embodiment constructs a full-link, in-depth defense security system from data generation, transmission, storage to access. Its core dynamic key generation technology makes each encryption process potentially unique, greatly increasing the difficulty of cracking and providing security for the most sensitive medical data of patients.

[0074] In one possible implementation, generating a first-round key for a symmetric encryption key and pre-converting the first-round key to obtain a pre-converted first-round key includes:

[0075] The first-round key for generating the symmetric encryption key is generated using the AES algorithm;

[0076] The first-round key is pre-converted into a decimal number to obtain the pre-converted first-round key.

[0077] In one possible implementation, multiple candidate keys are generated, evaluation values ​​of the candidate keys are obtained, and the best candidate key is obtained based on the evaluation values ​​of the candidate keys, including:

[0078] Generate multiple candidate keys. For the AES algorithm, the key length for each round is fixed, meaning that after conversion to decimal, it always falls within a fixed range. Therefore, multiple candidate keys can be randomly generated within this fixed range, and these multiple candidate keys can be randomly explored to determine the target key for the next round.

[0079] The Euclidean distance variance between the candidate key and the target key in the previous t rounds is obtained to obtain the evaluation value of the candidate key.

[0080] When obtaining the Euclidean distance variance between the candidate key and the target key of the previous t rounds, it is necessary to convert both the candidate key and the target key of the previous t rounds into their original bases (such as binary) before obtaining the Euclidean distance variance.

[0081] Based on the evaluation values ​​of all candidate keys, the candidate key with the highest evaluation value is determined as the best candidate key.

[0082] The higher the evaluation value of the candidate key, the weaker the correlation between the candidate key and the target key in the previous t rounds.

[0083] In one possible implementation, the single-point surround exploration includes:

[0084]

[0085]

[0086]

[0087]

[0088] in, Indicates the first t Round i Candidate keys Indicates the first i Candidate keys after a single-point surround exploration Indicates the first t The average of all candidate keys in the round. Indicates the best candidate key. Indicates the first single-point surrounding exploration factor. This indicates the second single-point surrounding exploration factor. Indicates the first t Round i The inertial weights of each candidate key. Indicates the first learning factor. Indicates the second learning factor. This represents the exploration angle coefficient, and , This represents the first constant term between [5, 10]. Represents pi (π). Represents a uniformly distributed random number between [0,1]. Represents the exploration quantity coefficient, and , This represents the exploration cycle coefficient between [0.5, 2]. Represents the corresponding parameter vectors The maximum value in, Represents the corresponding parameter vectors The maximum value in; This represents the maximum value of the inertia weight. This represents the minimum inertial weight. Indicates the first t Round i The evaluation value of each candidate key. This represents the evaluation value of the best candidate key. This represents the evaluation value of the worst candidate key. express The corresponding evaluation value.

[0089] Single-point exploration ensures in-depth exploration around the current optimal solution, preventing the loss of better neighboring solutions.

[0090] In one possible implementation, the combined random exploration includes

[0091]

[0092]

[0093]

[0094] in, Indicates the first m Candidate keys after a single-point surround exploration Indicates the first m Candidate keys after a combination of random explorations This represents the weighted key of the candidate keys after exploring all single-point perimeters. Indicates the best candidate key. Indicates the weighting coefficient. This indicates the total number of candidate keys.

[0095] Combinatorial stochastic exploration promotes genetic recombination among different high-quality solutions through information fusion, which helps to escape the local optimum trap.

[0096] In one possible implementation, the hop random exploration includes:

[0097]

[0098]

[0099] in, Indicates the first n Candidate keys after a combination of random explorations Indicates the first n Candidate keys after a random exploration of each hop. Indicates the best candidate key. This represents a random step size control factor between (0,1). Indicated by As the mean, with A Gaussian distributed random number with variance . Let represent the cosine function, and let represent pi.

[0100] Jump point random exploration provides global, large-scale random jump capability, ensuring the global convergence of the algorithm and avoiding getting trapped in local optima.

[0101] This multi-layered exploration strategy allows the key space to be searched fully and efficiently, and the generated keys are highly random, making them almost impossible to crack through exhaustive or statistical attacks.

[0102] In one possible implementation, based on the candidate keys obtained after exploration and the candidate keys themselves, an elite selection strategy is executed to obtain a selection key, and the best candidate key is updated based on the selection key to obtain the updated best candidate key, including:

[0103] Determine whether the evaluation value of the candidate key after exploration is greater than the evaluation value of its corresponding original candidate key. If so, determine the candidate key after exploration as the selection key; otherwise, determine the corresponding original candidate key as the selection key.

[0104] Determine the selection key with the largest evaluation value, and determine whether the evaluation value corresponding to the selection key with the largest evaluation value is greater than the evaluation value of the best candidate key. If so, the selection key with the largest evaluation value is used as the new best candidate key to obtain the updated best candidate key; otherwise, the original best candidate key is directly used as the updated best candidate key.

[0105] In one possible implementation, the preprocessed multimodal medical data is encrypted using a symmetric encryption key to obtain encrypted multimodal medical data, including:

[0106] Based on the symmetric encryption key, the preprocessed multimodal medical data is encrypted using the AES algorithm, and the symmetric encryption key is encrypted using the public key of the cloud server to obtain the encrypted multimodal medical data.

[0107] A hybrid mode of symmetric encryption (AES) and asymmetric encryption (RSA) is adopted. AES ensures fast encryption and decryption of massive amounts of medical data, while RSA solves the security problem of symmetric keys during transmission, thus balancing efficiency and security.

[0108] In one possible implementation, in response to an authorized diagnostic expert terminal initiating an access request to the cloud server, the diagnostic expert is authenticated, including:

[0109] In response to an authorized diagnostic expert terminal initiating an access request to the cloud server, the diagnostic expert undergoes at least one of facial recognition, iris recognition, and fingerprint recognition to obtain an authentication result; wherein the authentication result is either successful or unsuccessful.

[0110] By combining biometric identification (face, iris, fingerprint) for expert identity verification, it ensures that only legally authorized doctors can access the data, strengthening security protection at the end of the access chain and preventing identity theft.

[0111] This application's embodiments construct a comprehensive, layered security system encompassing data generation, transmission, storage, and access. Its core dynamic key generation technology ensures that each encryption process is potentially unique, significantly increasing the difficulty of cracking and providing financial-grade security for patients' most sensitive medical data.

[0112] This application's embodiments not only focus on security but also strive to optimize the diagnostic and treatment experience and efficiency. From encrypted data upload, cloud storage, expert verification, online diagnosis to automatic report generation and feedback, the entire process is highly integrated and automated. This reduces manual intervention, lowers the error rate, and shortens the cycle from data collection to report delivery. Experts do not need to worry about complex data encryption and decryption processes; they can directly access the data through simple biometric verification. The online interactive interface and automatic report generation function allow experts to focus on the diagnosis itself, rather than tedious paperwork, significantly improving diagnostic efficiency. Patients can obtain authoritative diagnostic reports more quickly, without long waits or travel to different hospitals. Data stored in chronological order also facilitates experts in tracking changes in the patient's condition, supporting continuous diagnosis and treatment.

[0113] The optimization mechanism of deep symmetric encryption algorithms possesses inherent robustness. Even if a candidate key performs poorly during the exploration process, the elite selection strategy and the group information sharing mechanism ensure that the algorithm converges towards the optimal direction as a whole, making it less susceptible to local noise interference. This method is not dependent on specific medical data types; whether it's images, gene sequences, or text medical records, it can be processed as multimodal medical data. The system architecture is clear, with decoupled modules (encryption, transmission, authentication, and diagnosis), facilitating future integration of new intelligent assisted diagnostic models, expansion to more types of biometric methods, or integration with different hospital information systems.

[0114] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0115] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, apparatuses, electronic devices, and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0116] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0117] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0118] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0119] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0120] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for remote diagnosis and report generation of multi-modal medical data, characterized in that, The method comprises the following steps: Collecting the corresponding multi-modal medical data of a patient, and preprocessing the multi-modal medical data to obtain preprocessed multi-modal medical data; Encrypting the preprocessed multi-modal medical data using a deep symmetric encryption algorithm to obtain encrypted multi-modal medical data; Transmitting the encrypted multi-modal medical data to a cloud server, and storing the encrypted multi-modal medical data in chronological order through the cloud server; In response to an access request initiated by an authorized diagnostic expert terminal to the cloud server, verifying the identity of the diagnostic expert, and allowing the diagnostic expert to access the encrypted multi-modal medical data if the verification is passed; Obtaining the diagnostic result input by the diagnostic expert through online human-computer interaction, filling the diagnostic result into a preset diagnostic report template to generate a patient diagnostic report, and returning the patient diagnostic report to the device corresponding to the patient; Encrypting the preprocessed multi-modal medical data using a deep symmetric encryption algorithm to obtain encrypted multi-modal medical data, comprising: Generating a first round key of the symmetric encryption key, pre-converting the first round key to obtain a pre-converted first round key, setting a key round number counter t to 1, and taking the pre-converted first round key as a target key of the tth round; Generating a plurality of candidate keys, obtaining evaluation values of the candidate keys, and obtaining a best candidate key according to the evaluation values of the candidate keys; According to the best candidate key, the candidate keys are sequentially subjected to single-point loop exploration, combined random exploration and jump point random exploration to obtain explored candidate keys; According to the explored candidate keys and the candidate keys, an elite selection strategy is executed to obtain a selection key, and the best candidate key is updated according to the selection key to obtain an updated best candidate key; The best candidate key is updated repeatedly for multiple times to determine a target key of the t+1th round, and it is judged whether the count value of the key round number counter is equal to the required number of rounds of the symmetric encryption key. If yes, all target keys are collectively taken as the symmetric encryption key, otherwise the count value of the key round number counter is increased by one, and the step of generating a plurality of candidate keys is returned; According to the symmetric encryption key, the preprocessed multi-modal medical data is encrypted to obtain encrypted multi-modal medical data; The single-point loop exploration comprises: wherein, denotes the t th candidate key, i denotes the th candidate key after the first single-point circumnavigation exploration, i denotes the th candidate key after the second single-point circumnavigation exploration, t denotes the mean of all candidate keys in the th round, denotes the best candidate key, denotes the first single-point circumnavigation exploration factor, denotes the second single-point circumnavigation exploration factor, t denotes the i th candidate key in the th round, denotes the first learning factor, denotes the second learning factor, denotes the exploration angle coefficient, and , denotes the first constant term between [5, 10], denotes the constant term of pi, denotes a random number between [0, 1], denotes the exploration quantity coefficient, and , denotes the exploration period coefficient between [0.5, 2]; denotes the maximum value in corresponding to all parameter vectors, denotes the maximum value in corresponding to all parameter vectors; denotes the maximum inertia weight, denotes the minimum inertia weight, t denotes the i th candidate key in the th round, denotes the best candidate key, denotes the worst candidate key, denotes the corresponding evaluation value; The combined random exploration comprises in, Indicates the first m Candidate keys after a single-point surround exploration Indicates the first m Candidate keys after a combination of random explorations This represents the weighted key of the candidate keys after exploring all single-point perimeters. Indicates the best candidate key. Indicates the weighting coefficient. Indicates the total number of candidate keys; The jump point random exploration comprises: in, Indicates the first n Candidate keys after a combination of random explorations Indicates the first n Candidate keys after a random exploration of each hop. Indicates the best candidate key. This represents a random step size control factor between (0,1). Indicated by As the mean, with A Gaussian distributed random number with variance . Let represent the cosine function, and let represent pi.

2. The remote diagnosis and report generation method of multi-modality medical data according to claim 1, characterized in that, Generating a first round key of the symmetric encryption key, and pre-converting the first round key to obtain a pre-converted first round key, comprising: Generating a first round key of the symmetric encryption key using the AES algorithm; The first round key is pre-converted into a decimal number to obtain a pre-converted first round key.

3. The remote diagnosis and report generation method of multi-modality medical data according to claim 1, characterized in that, Generating a plurality of candidate keys, obtaining evaluation values of the candidate keys, and obtaining a best candidate key according to the evaluation values of the candidate keys, comprising: Generating a plurality of candidate keys; Obtaining the Euclidean distance variance between the candidate key and the target key of the t rounds to obtain the evaluation value of the candidate key; According to the evaluation value of all candidate keys, the candidate key with the maximum evaluation value is determined as the best candidate key.

4. The remote diagnosis and report generation method of multi-modality medical data according to claim 1, characterized in that, According to the candidate key after the exploration and the candidate key, an elite selection strategy is performed to obtain a selected key, and the best candidate key is updated according to the selected key to obtain the best candidate key after the update, including: Determine whether the evaluation value of the candidate key after the exploration is greater than the evaluation value of the corresponding original candidate key, if yes, determine the candidate key after the exploration as the selected key, otherwise determine the corresponding original candidate key as the selected key; Determine the selected key with the maximum evaluation value, and determine whether the evaluation value corresponding to the selected key with the maximum evaluation value is greater than the evaluation value of the best candidate key, if yes, the selected key with the maximum evaluation value is taken as the new best candidate key to obtain the best candidate key after the update, otherwise the original best candidate key is directly taken as the best candidate key after the update.

5. The remote diagnosis and report generation method of multi-modal medical data according to claim 1, characterized in that, According to the symmetric encryption key, the preprocessed multi-modal medical data is encrypted to obtain encrypted multi-modal medical data, including: Based on the symmetric encryption key, the preprocessed multi-modal medical data is encrypted by using the AES algorithm, and the symmetric encryption key is encrypted by using the public key of the cloud server to obtain the encrypted multi-modal medical data.

6. The remote diagnosis and report generation method of multi-modality medical data according to claim 1, characterized in that, In response to the access request initiated by the authorized diagnostic expert terminal to the cloud server, the identity of the diagnostic expert is verified, including: In response to the access request initiated by the authorized diagnostic expert terminal to the cloud server, at least one of face recognition, iris recognition and fingerprint recognition is performed on the diagnostic expert to obtain an identity verification result; wherein the identity verification result is verification passed or verification failed.

Images