Safe access method for medical database
By using digital signature and comprehensive similarity analysis of user access behavior feature vectors in the medical database system, the problem of insufficient security of traditional password authentication is solved, timely identification and protection of suspicious access behavior is achieved, and the security of database access is improved.
Patent Information
- Application Number
- CN202510564964.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-30
- Publication Date
- 2025-08-15
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
During the existing medical database mutual access process, security is insufficient, traditional password authentication has the risk of leakage, and suspicious access behavior cannot be effectively identified.
Through the digital signature and comprehensive similarity analysis of user access behavior feature vectors, the central server judges the suspiciousness of access requests, uses feature extraction model and digital certificate verification, and uses weighted average similarity to judge the legitimacy of access behavior.
Improve the security during the inter-visiting process of medical databases, promptly detect suspicious access behaviors, and prevent potential security threats.
Smart Images
Figure CN120493276A_ABST
Abstract
Description
Technical field
[0001] The present invention belongs to the field of medical database management, and in particular relates to a safe access method for a medical database. [Background Technology]
[0002] In the prior art, different medical institutions (such as hospitals, medical research institutes, etc.) use their own databases to store data. Since different institutions need to access data from each other, each institution's database usually provides other authorized institutions with corresponding access interfaces and access rights.
[0003] Due to the sensitivity of medical data, it has high security requirements. Therefore, security checks are currently required for visitors. Existing security checks mainly rely on traditional technical means such as password authentication to ensure access security. However, due to the possibility of password leakage, the security of mutual access between existing medical databases is insufficient. [Summary of the invention]
[0004] In order to solve the above problems in the prior art, the present invention provides a method for securely accessing a medical database.
[0005] The technical solutions adopted in the present invention are as follows:
[0006] A method for securely accessing a medical database comprises the following steps:
[0007] Step 1: A user logs in to the medical database DB1 of the institution through a client, and submits an access request to another medical database through the medical database DB1; the access request includes a user ID, an ID of the target database to be accessed, and target data;
[0008] Step 2: The medical database DB1 obtains the user's access behavior feature vector V; the access behavior feature vector V is extracted by a feature extraction model based on the historical access behaviors of similar users of the user;
[0009] Step 3: The medical database DB1 digitally signs the access request and the access behavior feature vector V, and sends the access request, access behavior feature vector V, and digital signature to the central server;
[0010] Step 4: The central server verifies the digital signature. If the verification is successful, the central server sends the access request to all medical databases except the medical database DB1.
[0011] Step 5: For any medical database DB i (1≤i≤n), which obtains the user's historical access records based on the user identifier;
[0012] Step 6: Each medical database DB i , according to the user's historical access records and the access request, obtain the user's access behavior feature vector V for the medical database i , and V i Send to the central server;
[0013] Step 7: The central server calculates V i Similarity S with V i , and calculate the weighted average of n similarities as the comprehensive similarity P; assuming that the target database in the access request is DB2, DB3, ... ， DB k , other medical databases are DB k+1 , DB k+2 ,……,DB n , then the comprehensive similarity P is:
[0014]
[0015] W1, W2 and W3 are preset weights, W1>W2>W3;
[0016] Step 8: The central server determines whether the access request is suspicious based on the comprehensive similarity.
[0017] Furthermore, the feature extraction model obtains similar users of the user based on the user identifier, and summarizes and extracts corresponding access behavior features based on the access behaviors of all similar users over a period of time.
[0018] Furthermore, in step 3, the medical database uses the digital certificate of its corresponding medical institution to digitally sign the access request and the access behavior feature vector V to generate a triplet<Req,V,Sig> , where Req is the access request, Sig is the digital signature, and the medical database sends the triplet to the central server.
[0019] Furthermore, each medical database stores corresponding access records in a database log according to the actual access behavior of the user to the database; and extracts the historical access records of the user from the database log according to the user identifier in the access request.
[0020] Furthermore, in step 6, each medical database DB i The feature extraction model is used to extract the user's target data for the medical database DB according to the user's historical access records in the medical database and the target data to be accessed in the access request. i The access behavior feature vector Vi .
[0021] Furthermore, the similarity is calculated using a cosine similarity algorithm.
[0022] Furthermore, in step 8, the central server compares the comprehensive similarity with a predetermined threshold value, and if the comprehensive similarity is less than the predetermined threshold value, determines that the access request is a suspicious access behavior.
[0023] Furthermore, after discovering suspicious access behavior, the central server takes further processing measures for the suspicious access behavior.
[0024] Furthermore, the processing measures include issuing a warning to the administrator or terminating the user's access behavior.
[0025] The beneficial effect of the present invention is that suspicious access behaviors can be discovered in time during the mutual access process of the medical database, thereby greatly improving the security of the medical database access.
Brief Description of the Drawings
[0026] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application, but do not constitute an improper limitation of the present invention. In the drawings:
[0027] Figure 1 This is the main architecture diagram of the medical database system involved in the present invention. [Specific implementation method]
[0028] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The exemplary embodiments and descriptions are only used to explain the present invention but are not intended to limit the present invention.
[0029] See attached Figure 1 , which illustrates the main architecture of the entire medical database system involved in the present invention, and to which the method of the present invention is applied. The system includes multiple medical data servers, clients, and a central server. Each server, as well as the clients and medical data servers, can be interconnected and communicate with each other via a network.
[0030] Each medical data server runs a medical database. Generally speaking, different medical data servers belong to different medical institutions. It should be noted that each medical data server can be a single server or a server cluster, and this is not a limitation of the present invention. The medical database stores the medical data of the corresponding medical institution, which can include medical research data, patient medical records, and the institution's public medical basic data.
[0031] The client is a terminal device used by users to access medical databases. Each user belongs to a different medical institution and can use the client to log in to their institution's medical data server and access their institution's medical database. If necessary, users can request access to the medical databases of other medical institutions through their own institution's medical database. This login and access process can utilize traditional security authentication procedures, such as using an account and password.
[0032] The central server is the primary server that executes the method of the present invention. When a user accesses the medical database of another medical institution, their access behavior must pass a security check on the central server. By checking the user's database access behavior, the central server identifies suspicious access behavior and can then perform further suspicious behavior processing.
[0033] Based on the above system architecture, the medical database security access method of the present invention is described in detail below.
[0034] Step 1: A user logs in to the medical database DB1 of the institution through a client, and submits an access request to other medical databases through the medical database DB1.
[0035] Specifically, as mentioned above, each user belongs to a medical institution. Therefore, the user can use their client to access the institution's medical database. Specifically, the user connects to the institution's medical data server through their client, logs in, and accesses the medical database on that server. The process for logging in and accessing the institution's medical database follows existing access procedures and will not be detailed here.
[0036] After logging into the medical database of the institution to which the user belongs, if the user needs to access data in other medical databases, the user can submit an access request to the other medical databases through the medical database of the institution to which the user belongs.
[0037] The access request includes a user ID, an ID of the target database to be accessed, and the target data to be accessed. The user ID is a unique identifier used to identify the user and can be generated when the user registers. The ID of the target database can be pre-set, that is, a unique ID is pre-set for each medical database in the system and recorded on each medical data server. The user can select one or more database IDs from them as the target database IDs to be accessed. The target data can be represented by a data query statement or a data identifier, such as a database table name, database index, etc.
[0038] Step 2: The medical database DB1 obtains the user's access behavior feature vector V.
[0039] Specifically, the medical data server of the user's institution can generate an access behavior feature vector V for the user based on a pre-trained feature extraction model. Based on the user's identifier, the feature extraction model can obtain similar users to the user (including the user itself), summarize and extract corresponding access behavior features based on the access behavior of all similar users over a period of time, and represent these features as an access behavior feature vector V. Various feature extraction models for extracting user behavior features are available in the prior art and can all be applied to the present invention, without limitation thereto.
[0040] In short, an access behavior feature vector V can represent the access behavior characteristics of a type of users in a predetermined time period.
[0041] Step 3: The medical database DB1 digitally signs the access request and the access behavior feature vector V, and sends the access request, the access behavior feature vector V, and the digital signature to the central server.
[0042] Specifically, the medical database uses the digital certificate of its corresponding medical institution to digitally sign the access request and the access behavior feature vector V to generate a triplet<Req,V,Sig> Wherein, Req is the access request, Sig is the digital signature. The medical database sends the triplet to the central server.
[0043] Step 4: The central server verifies the digital signature. If the verification is successful, the central server sends the access request to all medical databases except the medical database DB1.
[0044] Specifically, the central server can verify the digital signature using the medical institution's digital certificate. If the signature verification passes, the access request is deemed legitimate. The central server then sends the access request to all other medical databases for subsequent security checks.
[0045] Assume that there are n medical databases DB1, DB2, ...DB n , the central server sends the access request to the DB i (2≤i≤n).
[0046] Step 5: For any medical database DB i (1≤i≤n), which obtains the user's historical access records based on the user identifier in the access request.
[0047] Specifically, each medical database stores access records in a database log based on the user's actual access behavior. Therefore, upon receiving the access request from the central server, each medical database can extract the user's historical access records from the database log based on the user identifier in the access request.
[0048] Step 6: Each medical database DB i , according to the user's historical access records and the access request, obtain the user's access behavior feature vector V for the medical database i , and access the behavioral feature vector V i Sent to the central server.
[0049] Specifically, similar to step 2, each medical database DB i Using the pre-trained feature extraction model, the user's target data for the medical database DB is extracted from the user's historical access records and the target data to be accessed in the access request. i The access behavior feature vector V i .
[0050] It should be noted that even if the target data to be accessed in the access request is in the medical database DB i It does not exist in the database, but it is also extracted together with the historical access records to extract features, which can effectively avoid the deceptive behavior of malicious visitors and effectively identify suspicious access.
[0051] Step 7: The central server calculates V i and V, obtaining n similarity values, and calculating a weighted average of the n similarity values as the comprehensive similarity.
[0052] Specifically, for V i and V, the similarity value S can be calculated i , so we can get n similarity values S1, S2, ..., S n The specific similarity algorithm may be any similarity algorithm known in the art, such as a cosine similarity algorithm, and the present invention does not limit this.
[0053] On this basis, the weighted average is calculated based on the three preset weights W1, W2, and W3, where W1>W2>W3. W1 corresponds to the medical database of the user's institution, W2 corresponds to the target database in the access request, and W3 corresponds to other medical databases in addition to the above databases.
[0054] Therefore, suppose the target database in the access request is DB2, DB3, ... ， DB k, other medical databases are DB k+1 , DB k+2 ,……,DB n , then the comprehensive similarity P is:
[0055]
[0056] Generally speaking, malicious actors will target key databases for malicious activity. Therefore, by adjusting weights, we can prioritize the user's organization database and target database, enabling more targeted identification of suspicious activity. Furthermore, by taking a weighted average of the similarities across all databases, we can also detect suspicious activity where malicious actors continually change their target databases.
[0057] Step 8: The central server determines whether the access request is suspicious based on the comprehensive similarity.
[0058] Specifically, the comprehensive similarity can measure the degree of deviation between user behavior and similar user behaviors. The smaller the comprehensive similarity, the greater the degree of deviation. Therefore, the central server can compare the comprehensive similarity with a predetermined threshold. If the comprehensive similarity is less than the predetermined threshold, it is determined that the access request is a suspicious access behavior.
[0059] After discovering suspicious access behavior, the central server can take further processing measures, such as issuing a warning to the administrator, or terminating the access behavior, etc. The focus of the present invention is not on how to handle it, which will not be described in detail.
[0060] If the access request is not suspicious, the central server may notify the target database to execute the access request.
[0061] Through the above steps, the present invention can timely discover suspicious access behaviors during the mutual access process of the medical database, thereby greatly improving the security of the medical database access.
[0062] The above description is only a preferred embodiment of the present invention. Therefore, any equivalent changes or modifications made according to the structure, characteristics and principles described in the scope of the patent application of the present invention are included in the scope of the patent application of the present invention.
Claims
1. A method for securely accessing a medical database, characterized in that: The following steps are involved: Step 1: A user logs in to the medical database DB1 of the institution through a client, and submits an access request to another medical database through the medical database DB1; the access request includes a user ID, an ID of the target database to be accessed, and target data; Step 2: The medical database DB1 obtains the user's access behavior feature vector V; the access behavior feature vector V is extracted by a feature extraction model based on the historical access behaviors of similar users of the user; Step 3: The medical database DB1 digitally signs the access request and the access behavior feature vector V, and sends the access request, access behavior feature vector V, and digital signature to the central server; Step 4: The central server verifies the digital signature. If the verification is successful, the central server sends the access request to all medical databases except the medical database DB1. Step 5: For any medical database DB i (1≤i≤n), which obtains the user's historical access records based on the user identifier; Step 6: Each medical database DB i , according to the user's historical access records and the access request, obtain the user's access behavior feature vector V for the medical database i , and V i Send to the central server; Step 7: The central server calculates V i Similarity S with V i , and calculate the weighted average of n similarities as the comprehensive similarity P; assuming that the target database in the access request is DB2, DB3, ... ， DB k , other medical databases are DB k+1 , DB k+2 ,……,DB n , then the comprehensive similarity P is: W1, W2 and W3 are preset weights, W1>W2>W3; Step 8: The central server determines whether the access request is suspicious based on the comprehensive similarity.
2. The method according to claim 1, characterized in that The feature extraction model obtains similar users of the user based on the user identifier, and summarizes and extracts corresponding access behavior features based on the access behaviors of all similar users over a period of time.
3. The method according to any one of claims 1 to 2, characterized in that In step 3, the medical database uses the digital certificate of the corresponding medical institution to digitally sign the access request and the access behavior feature vector V to generate a triplet<Req,V,Sig> , where Req is the access request, Sig is the digital signature, and the medical database sends the triplet to the central server.
4. The method according to any one of claims 1 to 3, characterized in that Each medical database stores corresponding access records in the database log according to the user's actual access behavior to the database; and extracts the user's historical access records from the database log according to the user identifier in the access request.
5. The method according to any one of claims 1 to 4, characterized in that In step 6, each medical database DB i The feature extraction model is used to extract the user's target data for the medical database DB according to the user's historical access records in the medical database and the target data to be accessed in the access request. i The access behavior feature vector V i .
6. The method according to claim 1, characterized in that The similarity is calculated using a cosine similarity algorithm.
7. The method according to claim 1, characterized in that In step 8, the central server compares the comprehensive similarity with a predetermined threshold value. If the comprehensive similarity is less than the predetermined threshold value, the central server determines that the access request is a suspicious access behavior.
8. The method according to claim 1, characterized in that After discovering suspicious access behavior, the central server takes further processing measures against the suspicious access behavior.
9. The method according to claim 8, characterized in that The action includes issuing a warning to an administrator.
10. The method according to claim 8, characterized in that The processing measures include terminating the user's access behavior.