A system and method for signing doctor-patient communication documents based on digital handwriting

Abstract

This application discloses a system for signing medical communication documents based on digital handwriting. The service governance platform Spring Cloud microservice framework is used as the core architecture layer of the business system. The access layer provides an access interface to access the access medium containing the electronic signature collection device; the access layer implements the intermediate authentication of the application layer and each business system of the smart hospital, and provides a service authentication entrance; the application layer provides signing task management, communication document management, signing document verification, and evidence information management services for the medical communication documents generated by each business system; the service layer converts the documents to be signed and unifies the format, calls the corresponding signing terminal to collect the signing handwriting according to the signing content, signing order, and signing position required by the signing task management to generate a signature data packet and merge the telegram to form a signed document, encrypts the signing data and signed documents and stores them; the resource layer supports the service layer, application layer service operation and signing server cluster resource equipment.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence and information security technology, and in particular to a system and method for generating doctor-patient communication documents using an electronic signature based on digital handwriting. Background Art

[0002] The electronic signature solution based on paperless signing of a large number of medical documents in hospitals has the following problems that need to be overcome: 1. The signing time of medical documents is not true, and a large number of medical documents are signed afterwards, which poses a risk of medical disputes; 2. The number of traditional paper medical documents signed in hospitals is large and difficult to preserve; 3. The authentication cost and time of paper medical documents are high; 4. The correlation between paper medical documents and medical records in the information system is low, requiring a lot of manpower for offline search and maintenance.

[0003] The large number of medical documents signed during medical communication urgently requires a safe and reliable electronic signature technology to effectively solve the generation and signing of electronic medical documents. Patients can also quickly obtain relevant medical information, such as informed consent forms, etc. Patients can sign safely online before receiving treatment, thereby simplifying the complex process of traditional paper signing. The electronic signature system needs to adopt advanced encryption technology to ensure the security of electronic signatures and document content. Make electronic documents authentic, complete and tamper-proof. The electronic signature system also includes an identity authentication mechanism to ensure that only authorized personnel can initiate the signature, and supports post-identification, which greatly reduces the risk of forgery and fraud.

[0004] Publication number: CN118280506A, name "A method and device for signing medical documents", obtain an electronic medical document to be signed by a patient as a medical document to be signed; display the medical document to be signed; record the patient's image data and voice data during the period of understanding the medical document to be signed; in response to detecting a trigger operation for a signature button, display a first signature area for signing the medical document to be signed; obtain the first signature data for the medical document to be signed through the first signature area; generate a signed electronic medical document based on the first signature data and the medical document to be signed; store the signed medical document to the medical service end. This application improves the signing efficiency of medical documents by signing the electronic medical document, and can also reduce the storage of paper medical documents.

[0005] The above-mentioned signature method relies on fixed hardware devices. It is necessary to build a signing system to generate electronic medical documents and then push them to the designated device for presentation and signature. It is not convenient for actual flexible use scenarios. In the actual treatment process, the signing of medical documents exists in various scattered departments. The same patient may need to sign the same document or different documents multiple times. The signing method that relies on fixed hardware devices is not convenient to use. In addition, the above-mentioned signature method requires the storage of image data and voice data during the signing period. Each document has multiple signature positions, so a large number of voice and video files will be generated, which requires high hard disk capacity and high cost.

[0006] Publication number: CN118214560A, name "Blockchain-based electronic prescription signature method and device", obtains a handwritten signature uploaded together with the electronic prescription, and authenticates the handwritten signature to generate authentication information of the handwritten signature; matches the doctor's private key based on the identity identifier, signs the electronic prescription with the private key, and obtains the electronic signature of the electronic prescription; distributes the prescription information and signature data of the signed electronic prescription through the blockchain network, and stores the authentication information of the electronic signature and the handwritten signature through the blockchain network, thereby improving the security of the electronic prescription and preventing the electronic signature of the electronic prescription from being tampered with.

[0007] The above signature method relies on the signatory to provide accurate identity identification and must obtain identity authentication before signing. During the authentication process, the private key needs to be prepared in advance (the form of the private key is unclear, it may be ukey). The operation is cumbersome and requires the signature data to be uploaded to the chain, which cannot meet the requirements of use in a closed network environment.

[0008] Publication number: CN118430727A, an electronic signature method based on CA authentication, a user with an administrator identity can manually configure a CA digital certificate authentication system as the main server, and other CA digital certificate authentication systems are configured with backup system identifications by the primary and backup management systems to become backup systems; a function is called to process the verification request sent by the electronic medical record system and send the request to the main system.

[0009] The above-mentioned signature method combines CA authentication with the electronic medical record system to request signatures on electronic documents in the electronic medical record system. The usage scenario is relatively rigid and cannot support the signature requirements of various treatment information systems during the medical treatment process.

[0010] There are many types of medical communication documents, and the contents and clauses that require patients' signatures and confirmation are complex, and the number of signatures is large. The signing equipment is different, involving multiple signatures in different environments and moods. The signatory's signature strokes, stroke order, pen grip, etc. may vary, which will result in large differences in handwriting style and characteristics of multiple signatures. A fixed handwriting data package cannot be generated for each signature data, and the generated and saved handwriting data packages may be quite different. The differences in signature data for medical communication documents in different treatment stages bring great difficulties to the subsequent verification and confirmation of the signed handwriting, and it is difficult to provide effective proof for the handling of medical disputes. Summary of the invention

[0011] The present invention aims at the problems that the existing technology for signing electronic documents and synthesizing signature files needs to rely on fixed equipment and requires private key authentication, which cannot meet the use of closed network environments, and the usage scenarios are fixed, and cannot support the signing of multiple and multi-style signatures issued by various business systems during the medical treatment process and the encryption of multiple signature information. A signing system and aspects for signing medical communication documents based on digital handwriting are proposed.

[0012] Based on the first aspect of the present application, a system for signing medical communication documents based on digital handwriting is proposed, including: an access layer, an application layer, a service layer, a resource layer, and an access layer; the access layer provides an access interface for accessing an access medium including an electronic signature collection device; the access layer is used to implement intermediate authentication between the application layer and various business systems of the smart hospital, and provide a service authentication entrance; the application layer is used to provide signing task management, communication document management, signing document verification, and evidence information management services for the medical communication documents generated by various business systems; the service layer is used to convert the documents to be signed and unify the format, call the corresponding signing terminal to collect the signing handwriting according to the signing content, signing order, and signing position required by the signing task management to generate a signature data packet and merge the electronic text to form a signing document, encrypt and store the signing data and signing document; the resource layer is used to support the service layer, application layer service operation and signing server cluster resource equipment.

[0013] Further optimization, the signing task management is responsible for creating, allocating, and tracking signing tasks, defining the signer, signing location, signing order, signing content, and setting the signing method; the communication document management module is responsible for classifying and archiving medical documents, providing medical communication documents to be signed, and synchronizing the information of unsigned and signed medical communication documents to the evidence information management module; the signing document verification module verifies the authenticity of all signatures on the signed medical communication documents, and submits the medical communication documents that have passed the verification to the evidence information management module for storage.

[0014] For further optimization, the service layer realizes the collection of electronic signature handwriting feature sequences, the processing, fusion, and deep feature extraction of handwriting feature data collected by multiple terminals, and the generation of signature data packets for communication documents in different links. The signature data packets are synthesized with the corresponding medical communication file telegrams to form signed documents, and the signature data packets and signed documents are encrypted and stored.

[0015] Further preferably, the service layer uses a neural network model to build a handwritten signature service, pre-processes multiple multi-style signatures collected at different links to extract multi-dimensional style feature representation vectors of the signatures, and the neural network model includes: a convolutional layer, a linear layer, an attention module, and a soft alignment module. The convolutional layer and the linear layer process the multi-dimensional style feature representation vector to obtain multiple style deep feature vectors of the deep layer of the handwriting. The attention module constructs a multi-style representation vector matrix based on the multi-dimensional style feature representation vector to calculate the attention weights of each style feature vector of the signature. The soft alignment module obtains the correlation of the corresponding style features of each dimension based on the deep feature vector to obtain the attention feature vector of each dimension, and the multi-style weighted feature vector is obtained by weighted calculation based on the attention feature vector and the attention weight, and the signature data packet is generated based on the weighted feature vector.

[0016] Further preferably, the signing method is set up including that the access layer converts the medical communication files of different formats received from various business systems into PDF or json / xml format files, and initiates a signing request to the signing task management module of the application layer. The signing task management module calls the handwriting signature service according to the customized signatory, signing terminal, and signing sequence to obtain the electronic signature handwriting data of the designated signing terminal to generate signature data, encrypt the signature data, and synthesize the encrypted signature data into a signature at a designated location of the designated medical communication file to generate a secure electronic file.

[0017] Further preferably, the generation of a secure electronic file includes a handwritten signature service unit calculating a hash value based on a picture of the document to be signed as auxiliary information for the signed document, using cumulative hash chain encryption to calculate hash values ​​of multiple signed documents and each signature data packet, accumulating a series of hash values ​​throughout the signing process to obtain a hash value for the signed document, superimposing the file auxiliary information and the hash value of the signed document as a signature value, and using the signature value to digitally sign the current signature data packet.

[0018] Further preferably, the generation of a secure electronic file includes, the handwritten signature service unit adding a signature domain to the formatted document to be signed according to the designated signing position, calculating the hash of the original text of the document to be signed, the hash of each signing information, the signing timestamp, and the identification cryptographic public key, encrypting a series of hash summaries of the entire signing process to obtain the signature value, using the original text hash value as the first field, the identification ID of the cryptographic public key as the second field, the signing timestamp as the third field, and the signature value as the fourth field to constitute the signing information, and synthesizing the signing information and the signature image into the signature domain.

[0019] Further preferably, the doctor-patient communication document is a to-be-signed document generated in association with the patient's ID number and / or medical card. The business system performs patient identity authentication and file transfer according to the interface specifications agreed upon by the access layer, pushes the to-be-signed document and patient identity information to the access layer, and the service layer starts the signing terminal signing page, renders the to-be-signed document as a fixed-format file, and starts signing task management.

[0020] Further preferably, handwriting collection obtains electronic signature handwriting data, and forms a signature data packet based on the handwriting data, which is [signatory information, handwriting point information, signature device information, auxiliary information]; the verification process includes digital certificate verification, timestamp verification and OCSP / CRL verification.

[0021] Based on the second aspect of the present application, a method for signing medical communication documents based on digital handwriting includes: the access layer provides an access interface to access an access medium including an electronic signature collection device; the access layer implements intermediate authentication between the application layer and each business system of the smart hospital, and provides a service authentication entry; the application layer provides signing task management, communication document management, signing document verification, and evidence information management services for the medical communication documents generated by each business system; the service layer converts the documents to be signed and unifies the format, calls the corresponding signing terminal to collect the signing handwriting according to the signing content, signing order, and signing position required by the signing task management to generate a signature data packet and merge the telegram to form a signing document, encrypts and stores the signing data and signing document; the resource layer supports the service layer and application layer service operations and the signing server cluster resource equipment.

[0022] Further preferably, the signing task management is responsible for creating, assigning, and tracking signing tasks, customizing the signer, signing order, signing content, and setting the signing method. The signing method setting includes the access layer converting the medical communication files of different formats received from various business systems into PDF or json / xml format files, and initiating a signing request to the signing task management module of the application layer. The signing task management module calls the handwritten signature service according to the customized signer, signing terminal, and signing order to obtain the electronic signature handwriting data of the designated signing terminal to generate signature data, encrypt the signature data, and synthesize the encrypted signature data into a signature at a designated location of the designated medical communication file to generate a secure electronic file.

[0023] Further preferably, the service layer uses a neural network model to build a handwritten signature service, pre-processes multiple multi-style signatures collected at different links to extract multi-dimensional style feature representation vectors of the signatures, and the neural network model includes: a convolutional layer, a linear layer, an attention module, and a soft alignment module. The convolutional layer and the linear layer process the multi-dimensional style feature representation vector to obtain multiple style deep feature vectors of the deep layer of the handwriting. The attention module constructs a multi-style representation vector matrix based on the multi-dimensional style feature representation vector to calculate the attention weights of each style feature vector of the signature. The soft alignment module obtains the correlation of the corresponding style features of each dimension based on the deep feature vector to obtain the attention feature vector of each dimension, and the multi-style weighted feature vector is obtained by weighted calculation based on the attention feature vector and the attention weight, and the signature data packet is generated based on the weighted feature vector.

[0024] Further preferably, each business system converts the doctor-patient communication documents into PDF format files or initiates a signing request to the patient signing system in json / xml format, unifies the format of the received documents, customizes the signatories and signing order, calls the signing service unit to generate signature data, synthesizes the doctor-patient communication documents in the unified format with electronic texts, and synthesizes the signatures at the specified location of the doctor-patient communication documents to generate a secure electronic file.

[0025] Further preferably, the generation of a secure electronic file includes a handwritten signature service unit calculating a hash value based on a picture of the document to be signed as auxiliary information for the signed document, using cumulative hash chain encryption to calculate hash values ​​of multiple signed documents and each signature data packet, accumulating a series of hash values ​​throughout the signing process to obtain a hash value for the signed document, superimposing the file auxiliary information and the hash value of the signed document as a signature value, and using the signature value to digitally sign the current signature data packet.

[0026] Further preferably, the generation of a secure electronic file includes, the handwritten signature service unit adding a signature domain to the formatted document to be signed according to the designated signing position, calculating the hash of the original text of the document to be signed, the hash of each signing information, the signing timestamp, and the identification cryptographic public key, encrypting a series of hash summaries of the entire signing process to obtain the signature value, using the original text hash value as the first field, the identification ID of the cryptographic public key as the second field, the signing timestamp as the third field, and the signature value as the fourth field to constitute the signing information, and synthesizing the signing information and the signature image into the signature domain.

[0027] The present invention extracts signature handwriting, synthesizes files, signs, and encrypts and stores multiple documents confirmed in each communication process between doctors and patients. It uses the uniqueness of electronic signature handwriting feature data to ensure the authenticity and certainty of each link in the doctor-patient communication process, and obtains secure electronic files through encryption processing of multi-signature data, providing powerful evidence for the handling of medical disputes. At the same time, it solves the problem that electronic signature data of different styles of signatures obtained under different backgrounds and emotional fluctuations during the diagnosis and treatment of patients can be deeply processed, and medical communication documents that can prove the true identity and intention of the signer can be stored for evidence.

[0028] The present invention is aimed at electronically signing informed documents for patients (or family members). Generate documents to be signed, configure signing process, sign notification, sign online, and generate documents to be signed. Configure signing process: execute signing tasks in sequence according to the process requirements of the initiator to initiate the signing, and the corresponding signatories complete the signing in sequence. Signing notification: push the signing notification according to the information of the signatory reserved by the initiator, and support signing according to the short link. Online signing: you can sign directly by accessing the various business system ends of the smart hospital, and also support signing after completing identity verification through SMS. Evidence storage process: the signed medical document signature confirmation file, signing end IP address, signing end identification code, timestamp, and business number generated during the signing process are associated through hashes, and synchronized to the intranet storage.

[0029] The present invention uses the handwritten original handwriting electronic signature technology to realize the signing of medical documents in the process of doctor-patient communication. The generated signature picture uses image steganography, and the signature picture is given new vitality. It contains the handwriting biometric information and telegram feature data of the signatory, which makes it beyond the scope of ordinary static signatures and establishes a unique and unreplicable connection between handwriting and telegram. Any attempt to illegally misappropriate this signature picture by copying, screenshots or taking photos will be easily verified as tampering due to the synthetic telegram. In addition, the combination of signature and telegram adopts identification password encryption technology to deeply integrate and encrypt the handwriting biometrics with the telegram. The patient signing system can be directly pushed to the signature board to sign, and the generated electronic file supports printing, which can cover the needs of traditional medical documents. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a flowchart of the signing of relevant medical documents during the doctor-patient communication process of the exemplary embodiment of this application;

[0031] Figure 2 This is a schematic diagram of the process of confirming and signing doctor-patient communication documents based on digital handwriting by the patient signing system in an exemplary embodiment of this application;

[0032] Figure 3 This is a schematic diagram of the association between the patient signing system and various business systems in the exemplary embodiment of this application. DETAILED DESCRIPTION

[0033] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings.

[0034] The present invention provides a method for signing relevant medical documents during the doctor-patient communication process based on a numerical signature of handwritten electronic handwriting. During the doctor-patient communication process, the patient signs the generated medical document in real time based on the digital handwriting through a signature pad, and the server merges the signed handwriting package and the signed document to generate a uniquely verifiable electronic document.

[0035] The main purpose of medical communication documents is to ensure that patients and their families fully understand key information such as the condition, treatment plan, possible risks and costs. By signing the document, patients and medical institutions can jointly confirm the treatment plan and clarify the responsibilities and risks of both parties. Therefore, the signed document must be tamper-proof and can serve as evidence to prove the signatory's intention in subsequent disputes. The digital handwriting signature is a unique data packet formed each time the patient (or family member) signs. It is biologically unique. The formation data and information in each handwriting package (signatory information, signing point information, device information) are comprehensively calculated to obtain a unique hash value, so that the original data can be traced back to the handwriting package and the signing process can be traced back to verify the source.

[0036] The electronic signature digital handwriting is based on the ability to more flexibly access mobile office devices, such as mobile phones, tablets, work PDAs and other smart handheld terminals. It supports software access to APIs, flexibly embeds signing capabilities into existing systems, and calls digital signature services through business drivers. It can also support office work on smart terminals, and there is no need to sign the same or different treatment period documents in a fixed place or on a fixed device.

[0037] The use of digital handwriting can achieve anonymous signatures by users, without the need to provide identity identification before signing. If there is any doubt about the signed document afterwards, the signed document can be verified, and the whole process supports the formation of a closed loop in the local area network, which does not rely on blockchain evidence storage. If there is a requirement for chaining, it can also be configured to be chained.

[0038] Digital handwriting can be used in any business system. It only needs to be connected to the signing control to achieve cross-platform digital signatures, reduce the binding relationship between systems, and expand the scope of electronic signature applications in medical scenarios.

[0039] The present invention proposes a patient signing system specifically for signing medical documents, which adopts the service governance core framework spring-cloud microservice architecture as the core architecture layer of the patient signing system, including: access layer, application layer, service layer, resource layer, and access layer. The access layer receives the medical documents to be signed pushed by each business system and unifies the file format, and pushes them to the patient signing terminal for interactive rendering and display (such as HTML5 page). The business system terminal that initiates the signing exchanges information with the signing terminal during the signing process, and stores and manages the data collected during the signing process. The service layer calls the signing service to collect handwriting, merge telegrams, encrypt management, store signature information, and encrypt and store them in the intranet.

[0040] The open signing interface with a unified integrated format uses a standard data format (supporting PDF, JSON, and XML), authenticates and authorizes the calling identity of the third-party business system through authentication mechanisms such as API Key and OAuth2.0, manages user authentication and interface access rights; returns HTTP status code and error information when a signing error or exception occurs; encrypts transmission data to ensure the security of the interface during transmission; prevents the interface from being maliciously attacked or overloaded, and implements current limiting and fuse strategies. This facilitates data exchange and analysis between different systems.

[0041] The electronic signature device collects signature handwriting feature data, obtains data such as strokes, pen sequence, pen pressure, and pen speed during the signing process, and the electronic signature device has a built-in high-speed processor to process the collected data in real time, perform preliminary data preprocessing and feature extraction, and generate a unique signature handwriting package. This handwriting package can not only be used for electronic signature verification, but also provide rich handwriting feature information to enhance the security and traceability of the signature. The handwriting data package consists of the device information during the collection process, signature handwriting point information (strokes, pen sequence, pen pressure, pen speed), and signatory information.

[0042] like Figure 1 Shown is a flowchart of the signing of relevant medical documents during the doctor-patient communication process in the exemplary embodiment of the present application.

[0043] The information system of smart hospitals includes: EMR (electronic medical record system), HIS (hospital information system), CIS (clinical information system), PACS (medical image archiving and communication system), LIS (laboratory information system), RIS (radiology information system), PEIS (physical examination information system), ORIS (operating room information system), HRP (hospital resource management system), HCRM (hospital customer relationship management system). When signing doctor-patient communication documents, it is necessary to establish links with each system to facilitate patients and their families to understand the doctor's diagnosis, medical experiments, examination results, etc., and to facilitate patients to confirm and sign after being informed.

[0044] When the patient's signature is required at a certain stage, any business system of the smart hospital (such as HIS, EMR, LIS) initiates a signing document request to the patient signature and transmits the document to be signed generated by the system. The patient signature system returns the signing link to the business system. The user opens the signing link through the business system and calls the signature and stamping module of the patient signature system to perform signing services for various business documents. The patient signature system returns the signed document to the business system that initiated the signing request.

[0045] The patient signing system includes: access layer, application layer, service layer, resource layer, and access layer.

[0046] The access layer provides an access interface for accessing the access medium including the electronic signature collection device. The access medium includes a computer, a signature pad (or tablet), and a signing SDK transition page.

[0047] The access layer is used to implement the intermediate authentication between the patient signing system and the business systems of the smart hospital, and provide external service authentication portals. It connects to the business systems of the smart hospital, such as HIS, EMR, and LIS systems, through the Internet or local interfaces, and is used to initiate access authentication services, signing initiation services, signing cancellation services, file download services, and other access services.

[0048] The application layer is used to provide signing services, signing task management, communication document management, signing document verification, and evidence information management. Signing task management includes: signing task creation, patient online signing, signing document viewing, and signing document downloading; patient document management includes: medical information registration, medical information modification, medical information export, and patient medical record viewing; signing document verification includes: signing document upload, signing document verification, verification result display, and signing information viewing; evidence information management includes: evidence information retrieval, evidence result download, evidence result display, and evidence report download.

[0049] The service layer is used to support the implementation of the application layer functions, including: handwriting signature service, handwriting collection service, fingerprint imprint service, handwriting recognition service, document conversion service, handwriting comparison service, file synthesis service, file verification service, and file storage service. Convert signed documents and unify the format, and call the signing service independently. Realize the collection of electronic signature handwriting feature sequences, the processing of handwriting feature data collected by multiple terminals, handwriting feature processing and deep feature extraction, generate user electronic signature data packets, synthesize signed data and medical communication document telegrams to form signed documents, encrypt and store signed data and signed documents, and later verify the authenticity of signatures and previous signatures during the communication process based on the encrypted signature data packet, signed documents, and original text.

[0050] The resource layer is used to support the operation of the service layer and application layer services and to sign the server cluster resource equipment. It includes: application servers, storage servers, and timestamp servers.

[0051] Signing task management module: responsible for allocating, creating, tracking and monitoring signing tasks. Control the patient document management module to receive documents to be signed, and determine the signing position and signing order according to the type and format of the document to be signed.

[0052] The signed document verification module receives the signed document and verifies it, and the evidence information management module receives and stores the document information that has passed the verification and preserves it for a long time.

[0053] Creating a signing task includes setting user-defined signatories and signing order, setting signing method, signing location, filing directory, filing watermark, notes and other information.

[0054] The patient document management module is responsible for collecting, organizing, classifying, archiving and reviewing medical documents from various business systems, including the storage and management of sensitive information such as medical records, examination reports, and surgical records. Classify and convert them into relevant format files, and provide corresponding doctor-patient communication documents to be signed for the signing task management module to assign relevant signing tasks; synchronize the unsigned and signed doctor-patient communication document information to the evidence information management module for backup and evidence storage.

[0055] The signing document verification module is responsible for verifying the electronic signature of the signed communication documents to ensure the authenticity, validity and legality of the signature. The verification process includes digital certificate verification, timestamp verification and OCSP / CRL verification. After the signing task is completed, the signing task management module controls the signing document verification module to receive the signed document, verify the electronic signature, submit the verified document information to the evidence information management module for long-term storage, and record the verification results.

[0056] The evidence information management module is responsible for storing and managing all signing information, verification results and document data related to doctor-patient communication documents. It provides data query, audit and traceability functions. It receives and stores information and data from signing task management, patient document management and signed document verification.

[0057] Signing task assignment and tracking

[0058] The signing task management module creates a signing task, extracts the doctor-patient communication documents to be signed from the patient document management module, and assigns them to the designated signatory. It is assigned to the designated signing terminal and the signatory and signing order are specified. A signing prompt is sent to the designated signing terminal, and the signing terminal interface prompts the signing operation. The patient document management module feeds back the signing progress to the signing task management module in real time.

[0059] Patient document management and filing

[0060] The patient document management module is responsible for collecting, organizing, and classifying patient medical documents from various business systems of the smart hospital, converting them into specified format files, and storing them in categories. When it is necessary to sign relevant doctor-patient communication documents, the doctor-patient communication documents to be signed are extracted, and the document information is provided to the signing task management module. After the signing is completed, the signed documents are archived and saved, and synchronized to the evidence information management module for backup.

[0061] Signature verification and authentication

[0062] After the signatory completes the signing, the document verification module verifies the signing result. The verification process includes digital certificate verification (to ensure the identity of the signer is authentic and valid), timestamp verification (to ensure the signature time is authentic and the signature is complete), and OCSP / CRL verification (to ensure the digital certificate has not been revoked at the time of signing). Documents that pass the verification will be considered valid signed documents and stored in the evidence information management module.

[0063] Evidence information management and query

[0064] The evidence information management module is responsible for storing all signature information, signature verification results and document data related to doctor-patient communication documents. The service provides users with information such as querying historical signature records, signature verification results and document data. The evidence information management and query module also provides data auditing and tracing functions to ensure the integrity and non-tamperability of data.

[0065] The present invention uses the patient signing system to sign the electronic medical document by handwriting the individual's name on the document that the doctor and patient have reached an agreement on. The server electronically merges and encrypts the signed document with the unique personal digital signature to meet the requirements of data security and post-verification.

[0066] Data encryption and storage

[0067] The patient document management module uses encryption algorithms (such as AES) to encrypt the medical document data that has completed the electronic signature storage. The signing task management module and the signed document verification module also use encrypted transmission and storage when transmitting and storing the signing information and verification results to ensure data security.

[0068] Permission control and access audit

[0069] The evidence information management module limits access rights to signed documents, authorizes access personnel and access time, and only allows authorized personnel to access relevant data at specified times. Implement the principle of minimum permissions to ensure that each authorized person only has the minimum permissions required to complete the work. Regularly review access records, promptly detect and handle abnormal access behaviors to ensure that data is not illegally accessed or tampered with.

[0070] Electronic signature and signature verification mechanism

[0071] The signing task management module defines the signatories and the signing order, and controls the access layer and the access layer to send the documents to be signed and the collected electronic signature handwriting feature sequence to the electronic signature synthesis signing platform or the electronic signature unit to sign and synthesize the medical communication documents.

[0072] The signed document verification module conducts strict electronic signature verification (including digital certificate verification, timestamp verification, and OCSP / CRL verification, etc.) on the signed documents to ensure the authenticity, validity, and legality of the signature. Documents that pass the verification will be regarded as valid signed documents and stored in the evidence information management module for subsequent query and use.

[0073] The signing of medical communication documents involves multiple confirmations, multiple signatures, and signing through different signing devices. The text and signatures may include multiple styles.

[0074] The service layer handwritten signature service unit of the present invention will process electronic signatures of different writing styles collected at different governance communication stages through different devices and at different times to extract multi-style weighted feature vectors, add the multi-style weighted feature vectors to the signature data packet and encrypt them.

[0075] The handwriting collection service unit collects and obtains multi-dimensional signature handwriting feature data signed multiple times by different signing devices, including geometric features of signature handwriting (such as pen tip coordinates, shape), dynamic features (such as pressure changes, dynamic time regularization, etc.), frequency domain features (such as rhythm features), time domain features (such as pen tip movement speed, acceleration) and pressure features (such as maximum pressure, average pressure). By extracting the feature vector data of different writing styles in the multi-dimensional signature features, multi-style feature vector data is obtained to construct a feature sequence or feature matrix as the only representation feature that reflects the user's writing signature style information. Further, the feature data can be compressed in time series to reduce storage space.

[0076] The handwritten signature service unit is constructed using a neural network model. The feature extraction module preprocesses multiple multi-style signatures to obtain a multi-dimensional style feature representation vector. The convolutional layer and linear layer in the neural network model process the multi-dimensional style feature representation vector to obtain multiple style deep feature vectors of the signature handwriting deep layer.

[0077] Among them, the multi-dimensional style features include: geometric features, dynamic features, frequency domain features, time domain features, and pressure features of electronic signature handwriting.

[0078] A convolutional neural network is used to build a neural network model, and the style feature vectors of each signature are extracted according to the style representation vector. The attention module in the convolutional neural network constructs a multi-style representation vector matrix based on the multi-dimensional style feature representation vector to calculate the attention weight of each style signature feature vector; the soft alignment module obtains the correlation of the style features of each dimension of the corresponding signature, obtains the attention feature vector of each dimension of the multi-style electronic signature, and performs weighted calculation based on each attention feature vector and the attention weight to obtain the multi-style weighted feature vector.

[0079] Among them, the attention module performs softmax operation on the signature feature vector to obtain the attention weight corresponding to each signature feature vector, which represents the correlation of the style features corresponding to each signature, and multiplies the attention weight with the feature vector of the corresponding signature to obtain the attention feature vector of the signature.

[0080] The softmax operation includes, for example, first performing the feature vector V ref Transpose and represent each dimension vector V separately test Multiply, according to the formula: Perform softmax operation to calculate the attention weight Atten of the signature feature vector. k Represents the dimension of the feature vector.

[0081] Evidence information management and long-term preservation

[0082] The evidence information management module is responsible for storing all signature information, signature verification results, document data and other information related to doctor-patient communication documents, and provides data query, audit and traceability functions to ensure data integrity and non-tamperability. A variety of backup and recovery strategies are also adopted to ensure the long-term preservation and availability of data.

[0083] The service governance platform Spring Cloud microservice framework is used as the core architecture of the business system to implement rich management functions including patient management, document management, equipment management, data storage and evidence collection, etc.

[0084] The patient signature system manages the basic information and documents of patients, including name, age, gender, contact information, etc. It stores and manages all medical documents signed by patients, and lists the signed documents in categories for easy access at any time. The signature terminal is managed and maintained through device management to ensure the normal operation and safety of the equipment.

[0085] Data storage and evidence collection: Save data including signature time, signer information, signature content, etc. during the signing process, and provide silent trace function so that it can be used as legal evidence when needed.

[0086] Receive documents with different contents and formats to be signed from various business systems (including HIS, EMR, LIS), and present them on the signing terminal through the front-end progressive framework.

[0087] Each business system converts the file into a PDF format file and initiates a signing request. If the business system does not convert the file format, the business system will initiate the signing content of the document to be signed in json / xml format. The business system accesses the access layer and sends the PDF format file or json / xml format file of the doctor-patient communication document to be signed to the patient signing system. The patient document management and archiving module converts the received file into a PDF format file and initiates a signing request.

[0088] When the doctor and the patient reach an agreement, the online document signing is initiated. This allows the patient to immediately electronically sign and confirm the generated medical documents, ensuring the data security and immutability of the medical documents.

[0089] Doctors initiate various medical document signing tasks to the patient signing system through the business system calling the signing service, and push the documents to be signed to the designated signing device (such as smart signature pads, mobile phones, tablets and other mobile terminal devices). The signing device collects the signed electronic documents and digital signature handwriting packages, and transmits them to the signing server to complete the electronic document fusion and evidence storage. When signing remotely on a mobile smart terminal, the patient signing system and the signing end use point-to-point transmission encryption and national encryption technology to achieve secure transmission of medical document signing information across network segments.

[0090] The format of the signed document pushed to the patient signing system must be PDF or json / xml structured data. The file is not identified during signing. The business system generates data or files based on actual business and then initiates a signing request. The access layer receives the document to be signed, the service layer converts the signed document and unifies the format, calls the signing server, and the application layer signing task management module creates a signing task and assigns a signing terminal and signing order. The document to be signed is pushed to the designated signing terminal, and the user (patient or family member) confirms the content of the document to be signed. The signing server receives the electronic signature handwriting feature data of the designated signing terminal in the signing order.

[0091] The signatory clicks the signing terminal to determine the signing position and can sign according to the document requirements, which may be a name or a confirmation word. The signing server echoes the signature handwriting synthesized to the designated position of the document to be signed based on the electronic signature handwriting feature data collected by the terminal.

[0092] The access layer converts the medical communication files of different formats received from various business systems into PDF or json / xml format files, and initiates a signing request to the signing task management module of the application layer. The signing task management module calls the handwritten signature service according to the customized signatory, signing terminal, and signing sequence to obtain the electronic signature handwriting data of the designated signing terminal to generate signature data, encrypt the signature data, and synthesize the encrypted signature data into a signature at the designated location of the designated medical communication file to generate a secure electronic file.

[0093] One of the embodiments of the present application for generating a secure electronic file includes: a handwritten signature service unit calculates a hash value based on a picture of the document to be signed as auxiliary information for the signed document, uses cumulative hash chain encryption to calculate hash values ​​of multiple signed documents and each signature data packet, accumulates a series of hash values ​​throughout the signing process to obtain a signed file hash value, superimposes the file auxiliary information and the signed file hash value as a signature value, and uses the signature value to digitally sign the current signature data packet.

[0094] One of the embodiments of the present application for generating a secure electronic file includes: a handwritten signature service unit adds a signature field to the to-be-signed format file according to a specified signature position, calculates the hash of the original text of the to-be-signed file, the hash of each signing information, the signing timestamp, and the identification cryptographic public key, encrypts a series of hash summaries of the entire signing process to obtain a signature value, sets the original text hash value as the first field, the identification ID of the cryptographic public key as the second field, the signing timestamp as the third field, and the signature value as the fourth field to constitute the signing information, and synthesizes the signing information and the signature image into the signature field.

[0095] like Figure 2 The figure shows a schematic diagram of the process of confirming and signing the doctor-patient communication documents based on digital handwriting by the patient signing system in an exemplary embodiment of the present application, including:

[0096] Basic information entry: Collect basic information including name, ID number, contact information, etc. and log it into the medical information business system.

[0097] Document preparation and signature: Generate the corresponding documents to be signed (such as informed consent, treatment plan, etc.) based on the patient's medical treatment situation through the ID card number, and push them to the patient signing system together with the collected patient information. Specifically,

[0098] The business system calls the specified service interface (performs identity authentication and file transfer according to the agreed interface specifications), starts the signing page of the patient signing system, and the signing page is rendered as a fixed-format file based on the received file, and a signing button pops up.

[0099] The patient signature system calls the electronic signature module to sign, supporting multiple signature methods such as handwritten signature collection and fingerprint collection. It records the signature timestamp, captures the signing process, and solidifies the biometric information to ensure the authenticity and non-tamperability of the signature.

[0100] The collected handwriting data is stored on the server side according to the handwriting package structure specification. The captured image can be used as auxiliary information to store the hash value of the image file as the auxiliary information of the signed PDF file (save the image hash in the PDF attachment data), and then the hash value of the signed file is calculated and encrypted for storage. n =Hash(Sum n -1+Hash n )Calculate the hash value of the signed file.

[0101] The collected handwriting data includes:

[0102]

[0103]

[0104] The synthesis module echoes the collected electronic signature handwriting time-series feature data as an image, synthesizes it locally into signed text and adds it to the designated signature position of the layout file. Add a signature field to the layout file in a unified format, calculate the electronic signature handwriting data based on the layout file to be signed and the electronic signature data package, and its electronic signature handwriting data may include: the original text hash of the document to be signed, calculate the original text hash based on the document to be signed pushed by the business system, obtain each signature handwriting package and the signing timestamp to calculate each signature information hash, encrypt to obtain a series of hash summaries of the entire signing process, calculate the signed file hash value based on the hash summary, and then calculate the signature value; obtain the cryptographic public key through the identification of the document to be signed, and the cryptographic public key can be used for subsequent signature verification. During the signature verification, the HASH value calculated based on the data after the signature information is removed after the document is signed is used as the original text HASH.

[0105] The synthesis module solidifies the original hash value in the first field of the signed information, the public key ID as the second field, the signing timestamp as the third field, and the signature value as the fourth field. The signed information and the signature image are synthesized into the signature field.

[0106] Document management and storage: After signing, the documents are automatically classified, filed and stored in the database. Patients and medical staff can query, access and download relevant documents through the system at any time.

[0107] Specifically, it can be classified according to business needs, such as partitioning and storing files in different departments, customizing word segmentation in the file management module, and storing corresponding files in corresponding groups for subsequent reference.

[0108] Data storage and evidence collection: Blockchain technology is used to store data to ensure data security and non-tamperability. When necessary, the system can provide data evidence collection services, issue evidence reports, and provide legal evidence for medical disputes.

[0109] According to the formula Sum n =Hash(Sum n -1+Hash n ) Calculate the hash value of the signed document and associate it with the business serial number for storage on the chain, so as to achieve unique comparison between the business and the document.

[0110] Document verification and review: Verify the signature of the signed document to verify the authenticity and validity of the signature. The system can review the signed document to ensure the compliance and integrity of the document.

[0111] When verifying the signature, first check whether the hash value of the file is consistent with the hash value calculated above. If they match, it means that the signed file has not been tampered with. Then, the signature information in the file is compared with the signature handwriting package stored in the corresponding server, and the signature information is verified to be consistent.

[0112] like Figure 3 The following is a schematic diagram of the association between the patient signature system and external systems / services (various business systems) in an exemplary embodiment of the present application. The patient signature system is intended to ensure that the patient's signature behavior during the medical process is safe, reliable and traceable. The association request process with external systems / services needs to ensure the accuracy of patient information, privacy protection and convenience of operation.

[0113] Each business system uploads a file to initiate a signing request. The signing service processes the file and configures the signing. It obtains the signature data of the signing terminal and sends it to the signing capability service to synthesize the file with the file to be signed. The signature is then stored locally or uploaded to the evidence storage center. At the same time, the signature on the synthesized file can be handwritten recognized and verified.

[0114] First, calculate the hash value of the file to be signed, then package it with the hash value of each signature information and calculate the hash value. Combine the hash of the file to be signed with each signature hash in the signing process to get the hash of the signed file. If backtracking is required, calculate the backtracking information of each signing node according to the encryption formula for verification.

[0115] Patient information registration and verification. When a patient uses the signature system for the first time, he or she needs to register in the system and enter personal information (such as name, ID number, contact information, etc.). The system verifies the authenticity of the patient information through external data sources (such as the data interface of the public security department).

[0116] Generate a signature key. The patient signature system signature verification module generates a unique signature key for each patient based on the registration information. The signature key is used for identity authentication and signature generation in the subsequent signing process.

[0117] The signing server assigns keys based on the signature information registered by each user, and when a user makes a request, the server assigns a random password as the current user's key.

[0118] The patient signing system needs to be associated with multiple external systems / services, such as hospital information system (HIS), electronic medical record system (EMR), patient relationship management system, etc. The patient signing system communicates with external systems / services through API interfaces or data exchange platforms to synchronize and share patient information.

[0119] The business system calls the specified service interface (which performs identity authentication and file transfer according to the agreed specifications) to retrieve the signing page of the patient signing system, and the signing page is rendered and displayed according to the received file. Click the screen position to start signing, and locate the signing position as the click position.

[0120] Signature request and verification. When a patient's signature is required, the business system sends a signature request to the patient signature system. After receiving the signature request, the patient signature system collects and obtains the electronic signature handwriting feature data according to the signatory and signing order defined by the signature task management module. The patient signature system can verify the signatory's identity information before signing. After verification, the collected signature handwriting feature data is sent to the electronic signature synthesis unit for synthesis of the signature file, and the signed file is returned to the business system.

[0121] When the identity of the signer needs to be verified, the patient signature system extracts the signature data and encrypted data package from the signature document of the relevant business system for verification to ensure the authenticity and validity of the signature.

[0122] The verification process includes matching based on the hash value of the electronic file. After a successful match, the corresponding signature data package is extracted and compared in sequence.

[0123] Recording and auditing, the patient signature system evidence information management module records each signature request, verification results and related operation logs. It provides an audit function that allows administrators to view and track any abnormalities or violations in the signature process.

[0124] The patient signing system keeps the signing process and the signed doctor-patient communication documents confidential. The main methods include:

[0125] The use of cumulative hash chain encryption storage technology has further enhanced the security of electronic data to ensure that it cannot be tampered with. At the same time, the use of desensitized HASH blockchain technology solidifies the evidence. In the process of confirming electronic medical documents, patients (or their relatives) need to sign the documents and related treatment information through the signing terminal, and encrypt the signed medical documents, signing information, and snapshot (or video) data of the signing process through the national cryptographic algorithm.

[0126] The signed document consists of the document to be signed and the signature information. The signature information will be included in the handwriting data package and the auxiliary information of the signed document for storage. The captured file will be encrypted by hash to obtain a unique hash value. For the signature file and signature handwriting data package in the medical process, the signature file is rendered in PDF format. The data package includes [user information, point information, device information, auxiliary information].

[0127] We calculate the hash of the signature handwriting data packets obtained by multiple signatures in the process of forming the doctor-patient communication document, and obtain a series of hash values ​​(Hash1, Hash2, Hash3, ..., Hashn). These hash values ​​are encrypted to obtain hash summaries (Sum1, Sum2, ..., Sumn). Among them, the hash summary of the nth document is calculated by the formula Sum n =Hash(Sum n -1+Hash n ) calculation, and calculate the signature value according to the hash summary. Use the signature value to sign the signed doctor-patient communication document and add the signature field.

[0128] Through chain encryption technology, point-to-point transmission encryption technology, and national password encryption technology, we cover the entire medical information confirmation business cycle of doctor-patient communication. In the Internet and hospital intranet environment, the security of data transmission is fully guaranteed, and the authenticity and effectiveness of the entire business are also confirmed. On this basis, we generate AOS files based on secure electronic files and store their hash summaries in the blockchain, thus achieving the goal of system security and easy use.

[0129] In addition, in each confirmation link of the treatment process, the parties (patients or their families) sign and confirm, and the evidence obtained meets the judicial requirements. At the same time, we use the SSL / TLS protocol to encrypt the data and verify the identities of both parties in the communication to ensure the security of information during transmission. In the process of signing electronic medical documents, we also record the biometric information of the signer in multiple dimensions and combine it with handwriting recognition technology to verify the authenticity of the signer's identity, thereby ensuring the authenticity and validity of the signed medical documents.

[0130] The signing terminal receives and displays the documents to be signed from the third-party business system. The patient (or relative) signs and confirms the electronic medical document by writing his or her personal name through the hardware device. The server integrates the signed document and the unique personal digital signature and performs encryption management. The present invention can be widely used in the process of doctor-patient communication. The patient can sign the generated medical documents immediately and with knowledge, which can reduce the paper consumables of the hospital, reduce the management and maintenance of paper documents by the hospital, and ensure the data security and non-tamperability of the medical documents.

[0131] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A system for signing doctor-patient communication documents based on digital handwriting, characterized in that: include: Access layer, application layer, service layer, resource layer, access layer; The access layer provides an access interface for accessing the access medium including the electronic signature collection device; The access layer is used to implement intermediate authentication between the application layer and various business systems of the smart hospital, and provide a service authentication entry; The application layer is used to provide signing task management, communication document management, signing document verification, and evidence information management services for the medical communication documents generated by various business systems; the service layer is used to convert the documents to be signed and unify the format, call the corresponding signing terminal to collect the signature handwriting according to the signing content, signing order, and signing position required by the signing task management, generate a signature data packet, and integrate the telegram to form a signed document, encrypt and store the signature data and signed documents; the resource layer is used to support the service layer, application layer service operation and signing server cluster resource equipment.

2. The system according to claim 1, characterized in that The signing task management is responsible for creating, assigning, and tracking signing tasks, defining the signer, signing location, signing order, signing content, and setting the signing method; the communication document management module is responsible for classifying and archiving medical documents, providing medical communication documents to be signed, and synchronizing the information of unsigned and signed medical communication documents to the evidence information management module; the signing document verification module verifies the authenticity of all signatures on the signed medical communication documents, and submits the medical communication documents that have passed the verification to the evidence information management module for storage.

3. The system according to claim 1, characterized in that The service layer realizes the collection of electronic signature handwriting feature sequences, the processing, fusion, and deep feature extraction of handwriting feature data collected by multiple terminals, and the generation of signature data packets for communication documents in different links. The signature data packets are combined with the corresponding medical communication file telegrams to form signed documents, and the signature data packets and signed documents are encrypted and stored.

4. The system according to claim 3, characterized in that The service layer uses a neural network model to build a handwritten signature service, and pre-processes multiple multi-style signatures collected at different stages to extract multi-dimensional style feature representation vectors of the signatures. The neural network model includes: convolutional layer, linear layer, attention module, and soft alignment module. The convolutional layer and linear layer process the multi-dimensional style feature representation vector to obtain multiple style deep feature vectors of the deep handwriting. The attention module constructs a multi-style representation vector matrix based on the multi-dimensional style feature representation vector to calculate the attention weights of each style feature vector of the signature. The soft alignment module obtains the correlation of the corresponding dimensional style features based on the deep feature vector to obtain the attention feature vector of each dimension. The multi-style weighted feature vector is obtained by weighted calculation based on the attention feature vector and the attention weight, and the signature data packet is generated based on the weighted feature vector.

5. The system according to claim 1, characterized in that The signing method described includes: the access layer converts the medical communication files of different formats received from various business systems into PDF or json / xml format files, and initiates a signing request to the signing task management module of the application layer. The signing task management module calls the handwritten signature service according to the customized signer, signing terminal, and signing sequence to obtain the electronic signature handwriting data of the designated signing terminal to generate signature data, encrypt the signature data, and synthesize the encrypted signature data into a signature at a designated location in the designated medical communication file to generate a secure electronic file.

6. The system according to claim 5, characterized in that The generation of a secure electronic file includes: a handwritten signature service unit calculates a hash value based on a picture of the document to be signed as auxiliary information for the signed document, uses cumulative hash chain encryption to calculate hash values ​​of multiple signed documents and each signature data packet, accumulates a series of hash values ​​throughout the signing process to obtain a hash value for the signed document, superimposes the document auxiliary information and the hash value of the signed document into a signature value, and uses the signature value to digitally sign the current signature data packet.

7. The system according to claim 5, characterized in that The generation of a secure electronic file includes the following steps: a handwritten signature service unit adds a signature field to the to-be-signed format file according to the designated signature position, calculates the hash of the original text of the to-be-signed file, the hash of each signature information, the signature timestamp, and the identification cryptographic public key, encrypts a series of hash summaries of the entire signing process to obtain a signature value, sets the original text hash value as the first field, the identification ID of the cryptographic public key as the second field, the signature timestamp as the third field, and the signature value as the fourth field to constitute the signature information, and synthesizes the signature information and the signature image into the signature field.

8. The system according to any one of claims 1 to 7, characterized in that: Medical communication documents are documents to be signed that are associated with the patient's ID number and / or medical card. The business system performs patient identity authentication and file transfer according to the interface specifications agreed upon by the access layer, and pushes the documents to be signed and the patient's identity information to the access layer. The service layer starts the signing terminal signing page, renders the documents to be signed as fixed-format files, and starts signing task management.

9. A method for signing doctor-patient communication documents based on digital handwriting, characterized in that: include: The access layer provides an access interface to access the access medium including the electronic signature collection device; the access layer implements the intermediate authentication between the application layer and each business system of the smart hospital, and provides a service authentication portal; the application layer provides signing task management, communication document management, signed document verification, and evidence information management services for the doctor-patient communication documents generated by each business system; the service layer converts the documents to be signed and unifies the format, and calls the corresponding signing terminal to collect the signature handwriting according to the signing content, signing order, and signing position required by the signing task management to generate a signature data packet and merge the telegram to form a signed document, and encrypts and stores the signature data and signed documents; The resource layer supports the service layer and application layer service operations and signs the server cluster resource equipment.

10. The method according to claim 9, characterized in that The signing task management is responsible for creating, assigning, and tracking signing tasks, customizing the signer, signing order, signing content, and setting the signing method. The signing method setting includes that the access layer converts the medical communication files of different formats received from various business systems into PDF or json / xml format files, and initiates a signing request to the signing task management module of the application layer. The signing task management module calls the handwritten signature service according to the customized signer, signing terminal, and signing order to obtain the electronic signature handwriting data of the designated signing terminal to generate signature data, encrypt the signature data, and synthesize the encrypted signature data into a signature at the designated location of the designated medical communication file to generate a secure electronic file.

11. The method according to claim 9, characterized in that The service layer uses a neural network model to build a handwritten signature service, and pre-processes multiple multi-style signatures collected at different stages to extract multi-dimensional style feature representation vectors of the signatures. The neural network model includes: convolutional layer, linear layer, attention module, and soft alignment module. The convolutional layer and linear layer process the multi-dimensional style feature representation vector to obtain multiple style deep feature vectors of the deep handwriting. The attention module constructs a multi-style representation vector matrix based on the multi-dimensional style feature representation vector to calculate the attention weights of each style feature vector of the signature. The soft alignment module obtains the correlation of the corresponding dimensional style features based on the deep feature vector to obtain the attention feature vector of each dimension. The multi-style weighted feature vector is obtained by weighted calculation based on the attention feature vector and the attention weight, and the signature data packet is generated based on the weighted feature vector.

12. The method according to any one of claims 9 to 11, characterized in that: Each business system converts the doctor-patient communication documents into PDF format files or initiates a signing request to the patient signing system in json / xml format, unifies the format of the received documents, customizes the signatories and signing order, calls the signing service unit to generate signature data, synthesizes the doctor-patient communication documents in the unified format, and synthesizes the signature in the specified location of the doctor-patient communication document to generate a secure electronic file.

13. The method according to claim 12, characterized in that The generation of a secure electronic file includes: a handwritten signature service unit calculates a hash value based on a picture of the document to be signed as auxiliary information for the signed document, uses cumulative hash chain encryption to calculate hash values ​​of multiple signed documents and each signature data packet, accumulates a series of hash values ​​throughout the signing process to obtain a hash value for the signed document, superimposes the document auxiliary information and the hash value of the signed document into a signature value, and uses the signature value to digitally sign the current signature data packet.

14. The method according to claim 12, characterized in that The generation of a secure electronic file includes the following steps: a handwritten signature service unit adds a signature field to the to-be-signed format file according to the designated signature position, calculates the hash of the original text of the to-be-signed file, the hash of each signature information, the signature timestamp, and the identification cryptographic public key, encrypts a series of hash summaries of the entire signing process to obtain a signature value, sets the original text hash value as the first field, the identification ID of the cryptographic public key as the second field, the signature timestamp as the third field, and the signature value as the fourth field to constitute the signature information, and synthesizes the signature information and the signature image into the signature field.

Images