Artificial intelligence-based craniocerebral injury data management method and device

Abstract

This application discloses an artificial intelligence-based method and apparatus for managing traumatic brain injury data, belonging to the field of data security technology. The method involves acquiring query requests for traumatic brain injury data and corresponding network traffic characteristic data, using an artificial intelligence model to analyze the network traffic characteristic data to determine network security status, and then deciding on data management actions based on the network security status. If secure, the request is responded to; otherwise, it is rejected and a warning is issued. When responding to a request, the target traumatic brain injury data is identified, and a target encryption key is generated using an aggregation-diffusion search algorithm. The encrypted data is then fed back to the user, and the target encryption key is transmitted using the user's public key. The method enhances network attack identification capabilities through an artificial intelligence model and generates a highly secure encryption key using an aggregation-diffusion search algorithm, thereby achieving secure management and transmission of traumatic brain injury data.

Description

Technical Field

[0001] This application belongs to the field of data security technology, specifically relating to an artificial intelligence-based method and device for managing craniocerebral injury data. Background Technology

[0002] Traumatic brain injury (TBI) data is highly sensitive personal medical privacy data, typically including patient imaging data (such as CT (Computed Tomography) images, MRI (Magnetic Resonance Imaging) images), clinical laboratory indicators, vital signs, and diagnostic results. With the rapid development of internet-based healthcare and telemedicine, the demand for cross-hospital and cross-regional sharing of TBI data is increasing. Existing TBI data management systems have the following security deficiencies: First, traditional access control relies heavily on static account passwords or simple firewall rules, making it difficult to effectively defend against covert SQL injection, distributed denial-of-service attacks, or advanced persistent threats disguised as legitimate requests. Second, regarding data transmission encryption, existing systems typically use fixed or simple random number generators to generate encryption keys. These key generation methods lack unpredictability and chaotic characteristics, making them vulnerable to brute-force attacks or side-channel attacks. Once the key space is reduced, attackers can decrypt and obtain the patient's private data. Summary of the Invention

[0003] The purpose of this application is to provide an artificial intelligence-based method and device for managing traumatic brain injury data, in order to solve the problems of weak anti-attack capability and insufficient security of encryption key generation mechanism in existing traumatic brain injury data management systems.

[0004] The first aspect of this application provides an artificial intelligence-based method for managing traumatic brain injury data, comprising: The system acquires query requests and network traffic characteristic data corresponding to traumatic brain injury data, and calls a pre-deployed artificial intelligence model to analyze the network traffic characteristic data to determine the network security situation. Based on the aforementioned network security situation, a decision is made regarding the user's current query request, and a data management action is determined; the data management action includes rejecting the query request and responding to the query request; If the data management action is to reject the query request, the user's query process will end, and an abnormal access warning message will be displayed. When the data management action is in response to a query request, the target traumatic brain injury data corresponding to the query request is determined; A target encryption key corresponding to the target traumatic brain injury data is generated using an aggregation-diffusion search algorithm, and the target encryption key is used to encrypt the target traumatic brain injury data to obtain the target encrypted data. The target encrypted data is fed back to the user, and the target encryption key is encrypted using the user's corresponding public key and transmitted to the user, thus completing the artificial intelligence-based management of traumatic brain injury data.

[0005] In one possible implementation, the pre-deployed artificial intelligence model is set as a CNN model or a CNN-LSTM model; The network traffic characteristic data is analyzed by invoking a pre-deployed artificial intelligence model to determine the network security situation, including: The network traffic characteristic data is used as input to a pre-deployed artificial intelligence model, and the analysis results output by the pre-deployed artificial intelligence model are obtained to determine the network security situation; the network security situation includes network security or network insecurity.

[0006] In one possible implementation, based on the network security situation, a decision is made regarding the user's current query request to determine data management actions, including: If the network security situation is secure, then the data management action is determined to be responding to the query request; If the network security situation is deemed insecure, the data management action will be to reject the query request.

[0007] In one possible implementation, a target encryption key corresponding to the target traumatic brain injury data is generated using an aggregation-diffusion search algorithm, including: Randomly generate multiple candidate encryption keys; Obtain the fitness value corresponding to each candidate encryption key; The candidate encryption key with the largest fitness value is determined as the optimal encryption key. Based on the optimal encryption key, a dual-objective aggregation strategy is used to quickly aggregate the candidate encryption keys to obtain the candidate encryption keys after rapid aggregation. A multi-point interpolation strategy is used to interpolate the candidate encryption keys after rapid aggregation to obtain interpolated candidate encryption keys. A multi-point diffusion strategy is used to perform a diffusion selection operation on the interpolated candidate encryption key to obtain the diffusion-selected candidate encryption key. Determine whether the number of iterations has reached the preset maximum number of iterations. If so, determine the target encryption key based on the candidate encryption keys selected after diffusion. Otherwise, return to the step of obtaining the fitness value based on the candidate encryption keys selected after diffusion.

[0008] In one possible implementation, a plurality of candidate encryption keys are randomly generated, including: For each round key of the symmetric encryption algorithm, determine the data space after converting each round key into a decimal number; In each round of key data space, a key value is randomly generated, and the key values ​​corresponding to all rounds are combined into a vector to obtain a candidate encryption key. Multiple different candidate secret keys are then obtained repeatedly.

[0009] In one possible implementation, the fitness value corresponding to each candidate encryption key is obtained as follows: in, Indicates the fitness value. Indicates the first weighted weight. Indicates the second weighting weight. Indicates the third weighting weight. Indicator of avalanche effect This indicates the encryption time of the candidate encryption key. Indicates the maximum allowed time. Indicates dissimilarity.

[0010] In one possible implementation, based on the optimal encryption key, a dual-objective aggregation strategy is used to quickly aggregate the candidate encryption keys, resulting in the following fast-aggregated candidate encryption keys: The first aggregation factor and the second aggregation factor are obtained as follows: in, This represents the first aggregation factor corresponding to the j-th candidate encryption key. Let represent the second aggregation factor corresponding to the j-th candidate encryption key, where j = 1, 2, ..., NP, and NP represents the total number of candidate encryption keys. This represents the aggregation radius parameter corresponding to the j-th candidate encryption key. This represents the aggregation angle parameter corresponding to the j-th candidate encryption key. Indicates the corresponding encryption keys among all candidate encryption keys Take the maximum value from the middle. Indicates the corresponding encryption keys among all candidate encryption keys Take the maximum value from the middle, and ; ; This represents the first constant term between (5, 10). This represents the second constant term between (5, 10). Represents pi (π). The first random number between (0,1) Represents a second random number between (0,1); Obtain the central encryption key corresponding to all candidate encryption keys; the parameter of each dimension of the central encryption key is the average value of the parameter of all candidate encryption keys in the same dimension; Using the optimal encryption key, the central encryption key, the first aggregation factor, and the second aggregation factor, the candidate encryption keys are quickly aggregated to obtain the following candidate encryption keys after rapid aggregation: in, Let j be the candidate encryption key in the t-th iteration. Let j be the candidate encryption key after the fast aggregation. The central encryption key, This is the optimal encryption key.

[0011] In one possible implementation, a multi-point interpolation strategy is used to interpolate the candidate encryption keys after rapid aggregation to obtain interpolated candidate encryption keys, including: For the candidate encryption keys after the rapid aggregation, two other candidate encryption keys after rapid aggregation are randomly matched to obtain a first random encryption key and a second random encryption key. Based on the first random encryption key and the second random encryption key, the candidate encryption keys after the fast aggregation are interpolated to obtain the interpolated candidate encryption keys as follows: in, Let d represent the d-th dimension key value of the candidate encryption key after the m-th fast aggregation, where d = 1, 2, ..., D, and D represents the number of key rounds required by the symmetric encryption algorithm. Let represent the d-th dimension key value of the candidate encryption key after the m-th interpolation process. This represents the d-th dimension of the first random encryption key. This represents the d-th dimension of the second random encryption key. This represents the fitness value of the first random encryption key. This represents the fitness value of the second random encryption key. This represents the fitness value of the candidate encryption key after the m-th fast aggregation.

[0012] In one possible implementation, a multi-point diffusion strategy is used to perform a diffusion selection operation on the interpolated candidate encryption key to obtain a diffusion-selected candidate encryption key, including: Obtain the fitness value of the candidate encryption key after interpolation, and determine the diffusion number corresponding to the candidate encryption key after interpolation based on the fitness value of the candidate encryption key after interpolation: in, This represents the fitness value of the candidate encryption key after the nth interpolation process. This represents the floor function. This represents the number of diffusions corresponding to the candidate encryption key after the nth interpolation process. This represents the maximum diffusion quantity, which is preset to an integer. This represents the minimum fitness value corresponding to all candidate encryption keys after interpolation. This represents the maximum fitness value corresponding to all candidate encryption keys after interpolation. indicates a non-zero decimal number, and is set to 0.001; NP represents the total number of candidate encryption keys; Based on the fitness value of the candidate encryption key after interpolation, the diffusion range corresponding to the candidate encryption key after interpolation is determined as follows: in, This represents the diffusion range corresponding to the candidate encryption key after the nth interpolation process. The fitness value representing the central encryption key; Based on the diffusion range corresponding to the candidate encryption key after the interpolation process, the diffusion candidate encryption key is generated as follows: in, Let d represent the d-th dimension key value of the candidate encryption key after the nth interpolation process in the t-th iteration. Let represent the d-th dimension key value of the diffused candidate encryption key corresponding to the nth interpolation-processed candidate encryption key. Indicates the weighting coefficient. This represents the maximum value corresponding to the weighting coefficient. This represents the fitness value of the optimal encryption key. This indicates that the mean is 0 and the variance is . The random values ​​are generated by a normal distribution, and T represents the preset maximum number of iterations; For any interpolated candidate encryption key, generate a random key. A total of 10 candidate encryption keys are selected for diffusion, and the candidate encryption key with the highest fitness value is selected as the candidate encryption key after diffusion selection.

[0013] A second aspect of this application provides an artificial intelligence-based traumatic brain injury data management device, comprising: The security analysis module is used to obtain query requests and corresponding network traffic feature data corresponding to traumatic brain injury data, and to call a pre-deployed artificial intelligence model to analyze the network traffic feature data to determine the network security situation. The management decision module is used to make decisions about the user's current query request based on the network security situation and determine the data management action; the data management action includes rejecting the query request and responding to the query request; The exception feedback module is used to terminate the user's query process and provide an exception access warning message when the data management action is to reject the query request. The data lookup module is used to determine the target traumatic brain injury data corresponding to the query request when the data management action is in response to the query request; The data encryption module is used to generate a target encryption key corresponding to the target traumatic brain injury data using an aggregation diffusion search algorithm, and to encrypt the target traumatic brain injury data using the target encryption key to obtain target encrypted data. The data feedback module is used to feed back the target encrypted data to the user, and to encrypt the target encryption key using the user's corresponding public key and transmit it to the user, thereby completing the artificial intelligence-based management of traumatic brain injury data.

[0014] The beneficial effects of this application are as follows: This application provides an artificial intelligence-based method and apparatus for managing traumatic brain injury data. The method acquires query requests for traumatic brain injury data and corresponding network traffic characteristic data, calls an artificial intelligence model to analyze the network traffic characteristic data to determine network security status, and then decides on data management actions based on the network security status. If secure, the request is responded to; otherwise, it is rejected and a warning is issued. When responding to a request, the target traumatic brain injury data is identified, and a target encryption key is generated using an aggregation-diffusion search algorithm. The encrypted data is then fed back to the user, and the target encryption key is transmitted using the user's public key. The method enhances network attack identification capabilities through an artificial intelligence model and generates a highly secure encryption key using an aggregation-diffusion search algorithm, thereby achieving secure management and transmission of traumatic brain injury data. Attached Figure Description

[0015] 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.

[0016] Figure 1 A flowchart illustrating an artificial intelligence-based method for managing traumatic brain injury data provided in this application; Figure 2 A schematic diagram of the structure of an artificial intelligence-based craniocerebral injury data management device provided in this application; Among them, 201-Security Analysis Module, 202-Management Decision Module, 203-Anomaly Feedback Module, 204-Data Search Module, 205-Data Encryption Module, and 206-Data Feedback Module. Detailed Implementation

[0017] 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.

[0018] The embodiments of this application are described in detail below with reference to the accompanying drawings.

[0019] like Figure 1 As shown in the figure, this application provides an artificial intelligence-based method for managing traumatic brain injury data, including: S101. Obtain the query request corresponding to the traumatic brain injury data and the corresponding network traffic characteristic data, and call the pre-deployed artificial intelligence model to analyze the network traffic characteristic data to determine the network security situation.

[0020] In one possible implementation, when a user initiates a query request for traumatic brain injury data (such as patient head CT images and diagnostic reports) to a medical data center via a terminal, network traffic characteristic data (such as packet length distribution, arrival time interval, TCP flag frequency, etc.) attached to the request can be captured simultaneously.

[0021] Preferably, the pre-deployed artificial intelligence model is set as a CNN (Convolutional Neural Network) model or a CNN-LSTM (Long Short-Term Memory) model. Taking the CNN-LSTM model as an example: the CNN convolutional layers are responsible for extracting local spatial features within a single data packet (such as abnormal payload length), while the LSTM is responsible for extracting temporal dependency features between data packet sequences (such as high-frequency short packet attack features). Network traffic feature data is input into the artificial intelligence model, and the output is the network security status, which is classified as either secure or insecure. It is worth noting that this artificial intelligence model can also be trained using existing network traffic datasets before use, such as the KDD Cup99 dataset.

[0022] This invention introduces an artificial intelligence model to perform real-time analysis of network traffic characteristic data before responding to data queries. By extracting the spatial and temporal series features of traffic through CNN or CNN-LSTM models, it can accurately identify network insecurity situations and directly reject requests before malicious traffic reaches the database, fundamentally avoiding the risk of malicious attacks consuming server resources or stealing data.

[0023] S102. Based on the network security situation, make a decision on the user's current query request and determine the data management action; the data management action includes rejecting the query request and responding to the query request. In one possible implementation, based on the network security situation, a decision is made regarding the user's current query request to determine data management actions, including: If the network security situation is secure, it indicates that the current traffic conforms to a normal user behavior pattern, and the data management action is determined to be responding to the query request; If the network security situation is insecure (e.g., SQL injection or DDoS attack characteristics are detected), indicating that the current user behavior is abnormal, the data management action is determined to be to reject the query request.

[0024] S103. If the data management action is to reject the query request, the user's query process is terminated, and an abnormal access warning message is sent. By directly disconnecting the connection and returning a warning message, the time the system is exposed to risk is minimized.

[0025] S104. If the data management action is in response to a query request, determine the target traumatic brain injury data corresponding to the query request; Based on the patient ID (Identity document) or examination number in the query request, the target traumatic brain injury data is located in the database, thereby enabling data retrieval.

[0026] S105. A target encryption key corresponding to the target traumatic brain injury data is generated using an aggregation diffusion search algorithm, and the target encryption key is used to encrypt the target traumatic brain injury data to obtain the target encrypted data. Existing technologies typically employ fixed or simple random number generators to generate encryption keys. The generated keys in each round exhibit a certain correlation, but this method lacks unpredictability and chaotic characteristics, making it vulnerable to brute-force or side-channel attacks. Once the key space is reduced, attackers can decrypt and obtain the patient's private data. Therefore, this application proposes an aggregation-diffusion search algorithm to generate the target encryption key corresponding to the target traumatic brain injury data. This algorithm possesses extremely high unpredictability and resistance to reverse engineering, thereby improving data security.

[0027] S106. Feed back the target encrypted data to the user, and encrypt the target encryption key using the user's corresponding public key and transmit it to the user, thereby completing the artificial intelligence-based management of traumatic brain injury data.

[0028] The target traumatic brain injury data is large in size, so a target encryption key (such as AES-256) is used for efficient symmetric encryption to obtain the target encrypted data. The target encryption key itself is small in size, so an asymmetric encryption algorithm (such as RSA) is used with a public key pre-uploaded by the user for encryption. Upon receiving the data, the user first decrypts the target encryption key using their private key, and then uses that key to decrypt the traumatic brain injury data, achieving a perfect balance between security and efficiency.

[0029] A high-strength, dynamically generated target encryption key is used to symmetrically encrypt the core data of traumatic brain injury. Simultaneously, the target encryption key is asymmetrically encrypted using the user's public key before transmission. Even if an attacker intercepts the data packet during network transmission, they cannot obtain the target encryption key without the user's private key, and therefore cannot reconstruct the traumatic brain injury data, thus achieving an end-to-end security loop.

[0030] In one possible implementation, a target encryption key corresponding to the target traumatic brain injury data is generated using an aggregation-diffusion search algorithm, including: Randomly generate multiple candidate encryption keys; Obtain the fitness value corresponding to each candidate encryption key; The candidate encryption key with the largest fitness value is determined as the optimal encryption key. Based on the optimal encryption key, a dual-objective aggregation strategy is used to quickly aggregate the candidate encryption keys to obtain the candidate encryption keys after rapid aggregation. A multi-point interpolation strategy is used to interpolate the candidate encryption keys after rapid aggregation to obtain interpolated candidate encryption keys. A multi-point diffusion strategy is used to perform a diffusion selection operation on the interpolated candidate encryption key to obtain the diffusion-selected candidate encryption key. Determine whether the number of iterations has reached the preset maximum number of iterations. If so, determine the target encryption key based on the candidate encryption keys selected after diffusion. Otherwise, return to the step of obtaining the fitness value based on the candidate encryption keys selected after diffusion.

[0031] The aggregation-diffusion search algorithm provided in this application uses a dual-objective aggregation strategy to enable candidate keys to converge quickly toward high fitness, a multi-point interpolation strategy to introduce nonlinear perturbations to prevent premature convergence, and a multi-point diffusion strategy to combine a normal distribution for local and global searches. The generated keys have extremely high unpredictability and resistance to reverse analysis.

[0032] In one possible implementation, a plurality of candidate encryption keys are randomly generated, including: For each round key of the symmetric encryption algorithm, determine the data space after converting each round key into a decimal number; In each round of key data space, a key value is randomly generated, and the key values ​​corresponding to all rounds are combined into a vector to obtain a candidate encryption key. Multiple different candidate secret keys are then obtained repeatedly.

[0033] For example, for the AES (Advanced Encryption Standard) algorithm, assuming 11 rounds of keys are required, the candidate secret key should contain 11 key values, and the number of bits corresponding to each round key is the same. That is to say, the converted decimal number has an upper and lower limit, and this upper and lower limit can form the data space. Since the length of each round key is fixed, the data space corresponding to all round keys is the same.

[0034] Optionally, to improve data computation efficiency, the data space can be set to (0,1]. After obtaining the key value, a certain round key can be obtained as the key value. (Upper limit - Lower limit) + Lower limit, then convert to the base corresponding to the symmetric encryption algorithm. It is worth noting that if, during the optimization process, decimals cause the length of the symmetric encryption key to differ from the original length after conversion, the key value in the candidate confidential key should be rounded up or down.

[0035] Existing dynamic keys often fall into local optima or lack sufficient pseudo-randomness. This invention designs an aggregation-diffusion search algorithm specifically tailored for cryptography. It ensures that the key population converges rapidly to high fitness through a dual-objective aggregation strategy, prevents premature convergence by introducing nonlinear perturbations through a multi-point interpolation strategy, and performs local and global searches by combining a multi-point diffusion strategy with a normal distribution. The keys generated by this algorithm have extremely high unpredictability and resistance to reverse analysis.

[0036] In one possible implementation, the fitness value corresponding to each candidate encryption key is obtained as follows: in, Indicates the fitness value. Indicates the first weighted weight. Indicates the second weighting weight. Indicates the third weighting weight. The avalanche effect index is indicated by observing the percentage of bits that change in the ciphertext when one bit is flipped in the key; the closer to 50%, the better. Indicates the encryption time of the candidate encryption key (ensuring real-time data query). Indicates the maximum allowed time. This indicates dissimilarity (the Hamming distance between the current key and historically generated keys, to prevent key duplication).

[0037] In generating the key, this invention does not employ a general optimization objective, but instead combines the Avalanche Effect (SAE) metric, encryption time, and Dissimilarity Index (DIS). This ensures that the final optimized target encryption key not only meets the requirements for real-time access to medical data in terms of encryption speed, but also possesses an extremely robust avalanche effect (small changes causing drastic changes in the ciphertext) and significant differences between keys, achieving commercial and even military-grade cryptographic security standards.

[0038] In one possible implementation, based on the optimal encryption key, a dual-objective aggregation strategy is used to quickly aggregate the candidate encryption keys, resulting in the following fast-aggregated candidate encryption keys: The first aggregation factor and the second aggregation factor are obtained as follows: in, This represents the first aggregation factor corresponding to the j-th candidate encryption key. Let represent the second aggregation factor corresponding to the j-th candidate encryption key, where j = 1, 2, ..., NP, and NP represents the total number of candidate encryption keys. This represents the aggregation radius parameter corresponding to the j-th candidate encryption key. This represents the aggregation angle parameter corresponding to the j-th candidate encryption key. Indicates the corresponding encryption keys among all candidate encryption keys Take the maximum value from the middle. Indicates the corresponding encryption keys among all candidate encryption keys Take the maximum value from the middle, and ; ; This represents the first constant term between (5, 10). This represents the second constant term between (5, 10). Represents pi (π). The first random number between (0,1) Represents a second random number between (0,1); Obtain the central encryption key corresponding to all candidate encryption keys; the parameter of each dimension of the central encryption key is the average value of the parameter of all candidate encryption keys in the same dimension; Using the optimal encryption key, the central encryption key, the first aggregation factor, and the second aggregation factor, the candidate encryption keys are quickly aggregated to obtain the following candidate encryption keys after rapid aggregation: in, Let j be the candidate encryption key in the t-th iteration. Let j be the candidate encryption key after the fast aggregation. The central encryption key, This is the optimal encryption key.

[0039] By introducing a first aggregation factor and a second aggregation factor, combined with the dynamic adjustment of the aggregation radius and aggregation angle parameters, candidate keys can quickly cluster towards the high-fitness regions where the optimal encryption key and the central encryption key reside, reducing the number of iterations and improving key generation efficiency. Random numbers are introduced into the aggregation factors to retain a certain degree of randomness during the aggregation process, avoiding local optima caused by excessive concentration of candidate keys and maintaining population diversity. Using the central encryption key (the mean of the same-dimensional parameters of all candidate keys) as a reference benchmark, and combined with the guiding role of the optimal encryption key, the aggregated keys are spatially closer to the global optimum, providing a high-quality initial population for subsequent interpolation and diffusion steps.

[0040] In one possible implementation, a multi-point interpolation strategy is used to interpolate the candidate encryption keys after rapid aggregation to obtain interpolated candidate encryption keys, including: For the candidate encryption keys after the rapid aggregation, two other candidate encryption keys after rapid aggregation are randomly matched to obtain a first random encryption key and a second random encryption key. Based on the first random encryption key and the second random encryption key, the candidate encryption keys after the fast aggregation are interpolated to obtain the interpolated candidate encryption keys as follows: in, Let d represent the d-th dimension key value of the candidate encryption key after the m-th fast aggregation, where d = 1, 2, ..., D, and D represents the number of key rounds required by the symmetric encryption algorithm. Let represent the d-th dimension key value of the candidate encryption key after the m-th interpolation process. This represents the d-th dimension of the first random encryption key. This represents the d-th dimension of the second random encryption key. This represents the fitness value of the first random encryption key. This represents the fitness value of the second random encryption key. This represents the fitness value of the candidate encryption key after the m-th fast aggregation.

[0041] By randomly matching two other rapidly aggregated candidate encryption keys (a first random encryption key and a second random encryption key) and combining their fitness values ​​for interpolation, nonlinear perturbations are introduced into the aggregated key population, effectively preventing the algorithm from getting trapped in local optima and improving the global optimization capability of key search. The interpolation formula generates new interpolated candidate encryption keys through a weighted combination of the dimension values ​​and fitness values ​​of different candidate encryption keys, enriching the diversity of the key population, avoiding excessive convergence of key values, and thus improving the anti-attack performance of the final generated target encryption key. Independent interpolation is performed on each round of keys required by the symmetric encryption algorithm, ensuring that the key values ​​of each round are fully optimized within their data space. This makes the generated key structure more cryptographically secure, providing higher-quality candidate keys for subsequent diffusion selection steps.

[0042] In one possible implementation, a multi-point diffusion strategy is used to perform a diffusion selection operation on the interpolated candidate encryption key to obtain a diffusion-selected candidate encryption key, including: Obtain the fitness value of the candidate encryption key after interpolation, and determine the diffusion number corresponding to the candidate encryption key after interpolation based on the fitness value of the candidate encryption key after interpolation: in, This represents the fitness value of the candidate encryption key after the nth interpolation process. This represents the floor function. This represents the number of diffusions corresponding to the candidate encryption key after the nth interpolation process. This represents the maximum diffusion quantity, which is preset to an integer. This represents the minimum fitness value corresponding to all candidate encryption keys after interpolation. This represents the maximum fitness value corresponding to all candidate encryption keys after interpolation. indicates a non-zero decimal number, and is set to 0.001; NP represents the total number of candidate encryption keys; Based on the fitness value of the candidate encryption key after interpolation, the diffusion range corresponding to the candidate encryption key after interpolation is determined as follows: in, This represents the diffusion range corresponding to the candidate encryption key after the nth interpolation process. The fitness value representing the central encryption key; Based on the diffusion range corresponding to the candidate encryption key after the interpolation process, the diffusion candidate encryption key is generated as follows: in, Let d represent the d-th dimension key value of the candidate encryption key after the nth interpolation process in the t-th iteration. Let represent the d-th dimension key value of the diffused candidate encryption key corresponding to the nth interpolation-processed candidate encryption key. Indicates the weighting coefficient. This represents the maximum value corresponding to the weighting coefficient. This represents the fitness value of the optimal encryption key. This indicates that the mean is 0 and the variance is . The random values ​​are generated by a normal distribution, and T represents the preset maximum number of iterations; For any interpolated candidate encryption key, generate a random key. A total of 10 candidate encryption keys are selected for diffusion, and the candidate encryption key with the highest fitness value is selected as the candidate encryption key after diffusion selection.

[0043] The diffusion quantity is dynamically determined based on the fitness values ​​of the interpolated candidate encryption keys. Keys with higher fitness values ​​receive more diffusion, while those with lower fitness values ​​receive fewer, achieving differentiated allocation of search resources. This ensures the stability of high-fitness keys while providing more optimization opportunities for low-fitness keys, thus improving the overall quality of the key population. A diffusion range formula correlates the fitness values ​​of the interpolated candidate encryption keys with those of the central encryption key, enabling dynamic adjustment of the local search range and balancing local fine-grained search with global exploration capabilities. When generating diffusion candidate encryption keys, randomness is introduced through the generation of random values ​​using a normal distribution. Simultaneously, a weighted coefficient is used to guide the diffusion process by using the fitness value of the optimal encryption key, ensuring that the diffusion process is both random enough to escape local optima and converges towards high-fitness regions, improving the global optimization capability and efficiency of key generation. Multiple diffusion candidate encryption keys are generated for each interpolated candidate encryption key, and the key with the highest fitness value is selected as the candidate encryption key after diffusion selection. This further filters out high-quality keys, improving the overall fitness level of the key population and providing a higher-quality initial population for subsequent iterative optimization.

[0044] In summary, generating the target encryption key corresponding to the target traumatic brain injury data through the aggregation-diffusion search algorithm effectively improves encryption security.

[0045] like Figure 2 As shown in the figure, this application provides an artificial intelligence-based traumatic brain injury data management device, including: The security analysis module 201 is used to obtain the query requests corresponding to the traumatic brain injury data and the corresponding network traffic characteristic data, and to call a pre-deployed artificial intelligence model to analyze the network traffic characteristic data to determine the network security situation. The management decision module 202 is used to make decisions on the user's current query request based on the network security situation and determine the data management action; the data management action includes rejecting the query request and responding to the query request; The exception feedback module 203 is used to terminate the user's query process and provide an exception access warning message when the data management action is to reject the query request. Data lookup module 204 is used to determine the target traumatic brain injury data corresponding to the query request when the data management action is in response to a query request; The data encryption module 205 is used to generate a target encryption key corresponding to the target traumatic brain injury data using an aggregation diffusion search algorithm, and to encrypt the target traumatic brain injury data using the target encryption key to obtain target encrypted data. The data feedback module 206 is used to feed back the target encrypted data to the user, and to encrypt the target encryption key using the user's corresponding public key and transmit it to the user, thereby completing the artificial intelligence-based management of traumatic brain injury data.

[0046] The artificial intelligence-based craniocerebral injury data management device described above can execute the technical solutions described in any of the above method embodiments. Its principles and beneficial effects are similar and will not be repeated here.

[0047] 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.

[0048] 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.

[0049] 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.

[0050] 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.

[0051] 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.

[0052] 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.

[0053] 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 managing traumatic brain injury data based on artificial intelligence, characterized in that, include: The system acquires query requests and network traffic characteristic data corresponding to traumatic brain injury data, and calls a pre-deployed artificial intelligence model to analyze the network traffic characteristic data to determine the network security situation. Based on the aforementioned network security situation, a decision is made regarding the user's current query request, and a data management action is determined; the data management action includes rejecting the query request and responding to the query request; If the data management action is to reject the query request, the user's query process will end, and an abnormal access warning message will be displayed. When the data management action is in response to a query request, the target traumatic brain injury data corresponding to the query request is determined; A target encryption key corresponding to the target traumatic brain injury data is generated using an aggregation-diffusion search algorithm, and the target encryption key is used to encrypt the target traumatic brain injury data to obtain the target encrypted data. The target encrypted data is fed back to the user, and the target encryption key is encrypted using the user's corresponding public key and transmitted to the user, thus completing the artificial intelligence-based management of traumatic brain injury data.

2. The artificial intelligence-based method for managing traumatic brain injury data according to claim 1, characterized in that, The pre-deployed artificial intelligence model is set to a CNN model or a CNN-LSTM model; The network traffic characteristic data is analyzed by invoking a pre-deployed artificial intelligence model to determine the network security situation, including: The network traffic characteristic data is used as input to a pre-deployed artificial intelligence model, and the analysis results output by the pre-deployed artificial intelligence model are obtained to determine the network security situation; the network security situation includes network security or network insecurity.

3. The artificial intelligence-based method for managing traumatic brain injury data according to claim 1, characterized in that, Based on the aforementioned network security situation, a decision is made regarding the user's current query request, determining data management actions, including: If the network security situation is secure, then the data management action is determined to be responding to the query request; If the network security situation is deemed insecure, the data management action will be to reject the query request.

4. The artificial intelligence-based method for managing traumatic brain injury data according to claim 1, characterized in that, The target encryption key corresponding to the target traumatic brain injury data is generated using an aggregation-diffusion search algorithm, including: Randomly generate multiple candidate encryption keys; Obtain the fitness value corresponding to each candidate encryption key; The candidate encryption key with the largest fitness value is determined as the optimal encryption key. Based on the optimal encryption key, a dual-objective aggregation strategy is used to quickly aggregate the candidate encryption keys to obtain the candidate encryption keys after rapid aggregation. A multi-point interpolation strategy is used to interpolate the candidate encryption keys after rapid aggregation to obtain interpolated candidate encryption keys. A multi-point diffusion strategy is used to perform a diffusion selection operation on the interpolated candidate encryption key to obtain the diffusion-selected candidate encryption key. Determine whether the number of iterations has reached the preset maximum number of iterations. If so, determine the target encryption key based on the candidate encryption keys selected after diffusion. Otherwise, return to the step of obtaining the fitness value based on the candidate encryption keys selected after diffusion.

5. The artificial intelligence-based method for managing traumatic brain injury data according to claim 4, characterized in that, Multiple candidate encryption keys are randomly generated, including: For each round key of the symmetric encryption algorithm, determine the data space after converting each round key into a decimal number; In each round of key data space, a key value is randomly generated, and the key values ​​corresponding to all rounds are combined into a vector to obtain a candidate encryption key. Multiple different candidate secret keys are then obtained repeatedly.

6. The artificial intelligence-based method for managing traumatic brain injury data according to claim 4, characterized in that, Obtain the fitness value for each candidate encryption key: in, Indicates the fitness value. Indicates the first weighted weight. Indicates the second weighting weight. Indicates the third weighting weight. Indicator of avalanche effect This indicates the encryption time of the candidate encryption key. Indicates the maximum allowed time. Indicates dissimilarity.

7. The artificial intelligence-based method for managing traumatic brain injury data according to claim 4, characterized in that, Based on the optimal encryption key, a dual-objective aggregation strategy is used to quickly aggregate the candidate encryption keys, resulting in the following fast-aggregated candidate encryption keys: The first aggregation factor and the second aggregation factor are obtained as follows: in, Indicates the first j The first aggregation factor corresponding to each candidate encryption key. Indicates the first j The second aggregation factor corresponding to each candidate encryption key. j =1,2,...,NP, where NP represents the total number of candidate encryption keys. Indicates the first j The aggregation radius parameter corresponding to each candidate encryption key. Indicates the first j The aggregation angle parameters corresponding to each candidate encryption key. Indicates the corresponding encryption keys among all candidate encryption keys Take the maximum value from the middle. Indicates the corresponding encryption keys among all candidate encryption keys Take the maximum value from the middle, and ; ; This represents the first constant term between (5, 10). This represents the second constant term between (5, 10). Represents pi (π). The first random number between (0,1) Represents a second random number between (0,1); Obtain the central encryption key corresponding to all candidate encryption keys; the parameter of each dimension of the central encryption key is the average value of the parameter of all candidate encryption keys in the same dimension; Using the optimal encryption key, the central encryption key, the first aggregation factor, and the second aggregation factor, the candidate encryption keys are quickly aggregated to obtain the following candidate encryption keys after rapid aggregation: in, For the first t During the nth iteration j One candidate encryption key, For the first j A number of candidate encryption keys after rapid aggregation. The central encryption key, This is the optimal encryption key.

8. The artificial intelligence-based method for managing traumatic brain injury data according to claim 7, characterized in that, A multi-point interpolation strategy is used to interpolate the candidate encryption keys after rapid aggregation, resulting in interpolated candidate encryption keys, including: For the candidate encryption keys after the rapid aggregation, two other candidate encryption keys after rapid aggregation are randomly matched to obtain a first random encryption key and a second random encryption key. Based on the first random encryption key and the second random encryption key, the candidate encryption keys after the fast aggregation are interpolated to obtain the interpolated candidate encryption keys as follows: in, Indicates the first m The first candidate encryption key after rapid aggregation d Dimensional key value, d =1,2,...,D, where D represents the number of key rounds required by the symmetric encryption algorithm. Indicates the first m The first interpolated candidate encryption key d Dimensional key value, The first random encryption key represents the... d Dimensional key value, The second random encryption key represents the... d Dimensional key value, This represents the fitness value of the first random encryption key. This represents the fitness value of the second random encryption key. Indicates the first m The fitness value of the candidate encryption keys after rapid aggregation.

9. The artificial intelligence-based method for managing traumatic brain injury data according to claim 8, characterized in that, A multi-point diffusion strategy is used to perform a diffusion selection operation on the interpolated candidate encryption keys to obtain diffusion-selected candidate encryption keys, including: Obtain the fitness value of the candidate encryption key after interpolation, and determine the diffusion number corresponding to the candidate encryption key after interpolation based on the fitness value of the candidate encryption key after interpolation: in, Indicates the first n The fitness value of the candidate encryption key after interpolation. This represents the floor function. Indicates the first n The number of diffusions corresponding to each candidate encryption key after interpolation. This represents the maximum diffusion quantity, which is preset to an integer. This represents the minimum fitness value corresponding to all candidate encryption keys after interpolation. This represents the maximum fitness value corresponding to all candidate encryption keys after interpolation. indicates a non-zero decimal number, and is set to 0.001; NP represents the total number of candidate encryption keys; Based on the fitness value of the candidate encryption key after interpolation, the diffusion range corresponding to the candidate encryption key after interpolation is determined as follows: in, Indicates the first n The diffusion range corresponding to each interpolated candidate encryption key. The fitness value representing the central encryption key; Based on the diffusion range corresponding to the candidate encryption key after the interpolation process, the diffusion candidate encryption key is generated as follows: in, Indicates the first t During the nth iteration n The first interpolated candidate encryption key d Dimensional key value, Indicates the first n The first interpolated candidate encryption key corresponds to the diffusion candidate encryption key of the first interpolated candidate encryption key. d Dimensional key value, Indicates the weighting coefficient. This represents the maximum value corresponding to the weighting coefficient. This represents the fitness value of the optimal encryption key. This indicates that the mean is 0 and the variance is . The random values ​​are generated by a normal distribution, and T represents the preset maximum number of iterations; For any interpolated candidate encryption key, generate a random key. A total of 10 candidate encryption keys are selected for diffusion, and the candidate encryption key with the highest fitness value is selected as the candidate encryption key after diffusion selection.

10. A data management device for traumatic brain injury based on artificial intelligence, characterized in that, include: The security analysis module is used to obtain query requests and corresponding network traffic feature data corresponding to traumatic brain injury data, and to call a pre-deployed artificial intelligence model to analyze the network traffic feature data to determine the network security situation. The management decision module is used to make decisions about the user's current query request based on the network security situation and determine the data management action; the data management action includes rejecting the query request and responding to the query request; The exception feedback module is used to terminate the user's query process and provide an exception access warning message when the data management action is to reject the query request. The data lookup module is used to determine the target traumatic brain injury data corresponding to the query request when the data management action is in response to the query request; The data encryption module is used to generate a target encryption key corresponding to the target traumatic brain injury data using an aggregation diffusion search algorithm, and to encrypt the target traumatic brain injury data using the target encryption key to obtain target encrypted data. The data feedback module is used to feed back the target encrypted data to the user, and to encrypt the target encryption key using the user's corresponding public key and transmit it to the user, thereby completing the artificial intelligence-based management of traumatic brain injury data.

Images