Archive management system and method thereof

Abstract

The invention discloses an archive management system and method, and relates to the technical field of intelligent archive management.The archive management method comprises the steps that paper archives are scanned into an electronic database, a number and a classification label are distributed to each paper archive, and a paper document text abstract is generated for each paper archive; searching the required paper archive according to the query condition, and displaying the detailed information of the required paper archive; modifying or deleting the paper archives needing to be modified, and recording the modifying or deleting time and operators at the same time; regularly copying the paper archives in the electronic database to an external storage device; and carrying out statistical analysis on the paper archives in the electronic database to generate a chart or a report. Therefore, the working efficiency and the archive management quality can be improved, the requirements of hospitals can be better met, and the archive management level is improved.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent archive management, and in particular to an archive management system and method thereof. Background Art

[0002] Archives management is an important task that involves the recording and preservation of information in various industries and fields. For example, in hospital application scenarios, traditional archives management methods mainly rely on the storage and retrieval of paper archives, but this method has many disadvantages, such as occupying a large space, being easily damaged, difficult to find, and easy to lose.

[0003] At present, some hospital archive management systems have been established and used to improve the work efficiency of hospitals and protect the privacy and safety of patients. However, there are also some problems with the existing archive management systems. For example, the system functions are not perfect enough to meet the needs of various departments and levels of the hospital, such as the inability to analyze archives, resulting in poor quality of archive management.

[0004] Therefore, an optimized archive management system and method thereof are desired. Summary of the invention

[0005] The present application provides an archive management system and method thereof, which scans paper archives into an electronic database, assigns a number and classification label to each of the paper archives, and generates a paper document text summary for each of the paper archives; searches for the required paper archives according to the query conditions, and displays the detailed information of the required paper archives; modifies or deletes the paper archives that need to be modified, and records the time and operator of the modification or deletion; regularly copies the paper archives in the electronic database to an external storage device; performs statistical analysis on the paper archives in the electronic database, and generates charts or reports. In this way, work efficiency and archive management quality can be improved, thereby better meeting the needs of the hospital and improving the level of archive management.

[0006] The present application also provides a file management system, which includes: The file entry module is used to scan paper files into the electronic database, assign a number and classification label to each paper file, and generate a paper document text summary for each paper file; The file query module is used to search for the required paper files according to the query conditions and display the detailed information of the required paper files; The file modification module is used to modify or delete the paper files that need to be modified, and record the time of modification or deletion and the operator; An archive backup module, used to regularly copy paper archives in the electronic database to an external storage device; An archive statistics module is used to perform statistical analysis on paper archives in the electronic database and generate charts or reports; Among them, the archive entry module includes: a scanned file acquisition unit, used to acquire the scanned file of the paper archive to be analyzed; a text recognition unit, used to use text recognition technology to perform text recognition on the scanned file to obtain a paper archive text description; a semantic feature extraction unit, used to extract the semantic features of the paper archive text description to obtain a sequence of semantic encoding feature vectors of paper document text segment descriptions; an essential semantic information extraction unit, used to pass the sequence of semantic encoding feature vectors of paper document text segment descriptions through an essential semantic information extraction network to obtain a paper document text essential semantic information encoding feature vector; a text summary generation unit, used to obtain the paper document text summary based on the paper document text essential semantic information encoding feature vector.

[0007] In the above-mentioned archive management system, the semantic feature extraction unit includes: a data preprocessing subunit, used to perform data preprocessing on the paper archive text description to obtain a sequence of paper document text segment descriptions; a semantic coding analysis subunit, used to perform semantic coding analysis on the sequence of paper document text segment descriptions to obtain a sequence of semantic coding feature vectors of the paper document text segment descriptions.

[0008] In the above-mentioned archive management system, the data preprocessing subunit is used to: segment the paper archive text description to obtain a sequence of paper document text segment descriptions.

[0009] In the above-mentioned archive management system, the semantic coding analysis subunit is used to: pass each paper document text segment description in the sequence of the paper document text segment description through a text convolutional neural network model including a word embedding layer to obtain a sequence of semantic coding feature vectors of the paper document text segment description.

[0010] In the above-mentioned archive management system, the semantic coding analysis subunit includes: a word segmentation secondary subunit, which is used to perform word segmentation processing on each paper document text segment description in the sequence of the paper document text segment description to convert each paper document text segment description in the sequence of the paper document text segment description into a word sequence composed of multiple words; an embedded coding secondary subunit, which is used to use the word embedding layer of the text convolutional neural network model including the word embedding layer to map each word in the word sequence to a word vector to obtain a sequence of word vectors; a context coding secondary subunit, which is used to use the text convolutional neural network model including the word embedding layer to semantically encode the sequence of word vectors to obtain a sequence of semantic coding feature vectors of the paper document text segment description.

[0011] In the above-mentioned archive management system, the essential semantic information extraction unit is used to: process the sequence of semantic coding feature vectors describing the paper document text segment using the following essential semantic information extraction formula to obtain the paper document text essential semantic information coding feature vector; wherein the essential semantic information extraction formula is: ; ; in, is the first in the sequence of semantic coding feature vectors describing the paper document text segment A paper document text segment describes the semantic encoding feature vector, is the first in the sequence of semantic coding feature vectors describing the paper document text segment A paper document text segment describes the semantic encoding feature vector, represents the 1-norm of the eigenvector, The length of the sequence of semantic encoding feature vectors describing the paper document text segment - 1, A representation of a sequence of semantic encoding feature vectors describing the paper document text segment, represents the semantic feature difference coefficient, represents the natural exponential function operation, represents the total number of difference coefficients of the semantic features, A feature vector is encoded for the essential semantic information of the paper document text.

[0012] In the above-mentioned archive management system, the text summary generation unit includes: an optimization subunit, which is used to optimize the encoding feature vector of the essential semantic information of the paper document text to obtain an optimized encoding feature vector of the essential semantic information of the paper document text; and an archive summary generation subunit, which is used to pass the optimized encoding feature vector of the essential semantic information of the paper document text through an archive summary generator based on AIGC to obtain the paper document text summary.

[0013] The present application also provides a file management method, which includes: Scanning paper archives into an electronic database, assigning a number and classification label to each paper archive, and generating a paper document text summary for each paper archive; According to the query conditions, the required paper archives are searched and the detailed information of the required paper archives is displayed; Modify or delete paper files that need to be modified, and record the time of modification or deletion and the operator; Periodically copy the paper records in the electronic database to an external storage device; Conduct statistical analysis on the paper files in the electronic database and generate charts or reports; Among them, the paper archives are scanned into an electronic database, a number and a classification label are assigned to each of the paper archives, and a paper document text summary is generated for each of the paper archives, including: obtaining a scanned file of the paper archive to be analyzed; using text recognition technology to perform text recognition on the scanned file to obtain a paper archive text description; extracting semantic features of the paper archive text description to obtain a sequence of semantic coding feature vectors of paper document text segment descriptions; passing the sequence of semantic coding feature vectors of paper document text segment descriptions through an essential semantic information extraction network to obtain a paper document text essential semantic information coding feature vector; and obtaining the paper document text summary based on the paper document text essential semantic information coding feature vector.

[0014] Compared with the existing technology, the archive management system and method provided by the present application can use text recognition technology and natural language processing technology based on deep learning to perform semantic analysis on paper archives, thereby intelligently generating text summaries of paper documents to improve work efficiency and archive management quality, thereby better meeting the needs of hospitals and improving the level of archive management. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings: Figure 1 A block diagram of an archive management system provided in an embodiment of the present application.

[0016] Figure 2 This is a flowchart of a file management method provided in an embodiment of the present application.

[0017] Figure 3 A schematic diagram of the system architecture of a file management method provided in an embodiment of the present application.

[0018] Figure 4 This is an application scenario diagram of an archive management system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the embodiments of the present application more clear, the embodiments of the present application are further described in detail below in conjunction with the accompanying drawings. Here, the illustrative embodiments of the present application and their descriptions are used to explain the present application, but are not intended to limit the present application.

[0020] Unless otherwise specified, all technical and scientific terms used in the embodiments of the present application have the same meaning as those commonly understood by those skilled in the art of the present application. The terms used in this application are only for the purpose of describing specific embodiments and are not intended to limit the scope of this application.

[0021] In the embodiments of the present application, it should be noted that, unless otherwise specified and limited, the term "connection" should be understood in a broad sense. For example, it can be an electrical connection or a connection between two components. It can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0022] It should be noted that the terms "first\second\third" involved in the embodiments of the present application are merely used to distinguish similar objects, and do not represent a specific order for the objects. It is understandable that the "first\second\third" can be interchanged with a specific order or sequence where permitted. It should be understood that the objects distinguished by "first\second\third" can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein.

[0023] The importance of archive management in hospitals is self-evident. Traditional paper archive management methods have many problems, such as taking up a lot of space, being easily damaged, difficult to find, and easy to lose. Therefore, many hospitals have begun to adopt electronic archive management systems to solve these problems, improve work efficiency, and protect patient privacy and safety.

[0024] Electronic records management systems can bring many benefits, including but not limited to: 1. Save space: Electronic files do not take up actual space, which can greatly reduce the hospital's file storage costs.

[0025] 2. Safe and reliable: Electronic files can be encrypted and permission-managed to ensure patient privacy and file security.

[0026] 3. Easy to find: Through keyword search and classification tags, you can quickly and accurately find the required archival information.

[0027] 4. Long-term preservation: Electronic archives can be more easily preserved and backed up for a long time to avoid information loss.

[0028] However, the existing electronic archive management system also has some problems, such as incomplete functions, which cannot meet the needs of various departments and levels of the hospital, resulting in poor quality of archive management. In order to solve these problems, the following improvement measures can be considered: Customization function, customize the archive management system according to the actual needs of the hospital to meet the needs of different departments and levels. Data analysis function, add data analysis function to help hospitals better use archive information for scientific research and decision-making analysis. User training, train hospital staff on the use of the system to ensure that they can make full use of the functions of the archive management system.

[0029] However, there are some problems with electronic archive management systems. Some electronic archive management systems may lack the functions required by specific departments or user groups. For example, they may lack specific types of data analysis functions, collaboration functions, or customized functions required by specific departments. Some systems may have problems such as complex user interfaces and cumbersome operations, which may lead to low work efficiency and employees' reluctance to accept the system. In some cases, different systems may not be well integrated, which may lead to information silos, making cross-departmental or cross-system information sharing and collaboration difficult. Electronic archive management systems need to meet strict privacy and security standards to protect sensitive information of patients and hospitals. There is a risk that security vulnerabilities may lead to information leakage or tampering. Incomplete, inaccurate or outdated data may have a negative impact on the decision-making and work of the hospital. Therefore, the system needs to ensure the accuracy and integrity of the data. Some systems may face stability and performance issues due to aging technology or lack of timely maintenance and updates.

[0030] These problems may affect the hospital's work efficiency, the security of patient information and the overall management level of the hospital. In this application, an optimized archive management system and method thereof are provided.

[0031] In one embodiment of the present application, Figure 1 Schematic diagram of a file management system provided in an embodiment of the present application. Figure 1As shown, the archive management system 100 according to the embodiment of the present application includes: an archive entry module 110, which is used to scan paper archives into an electronic database, assign a number and a classification label to each of the paper archives, and generate a paper document text summary for each of the paper archives; an archive query module 120, which is used to find the required paper archives according to the query conditions, and display the detailed information of the required paper archives; an archive modification module 130, which is used to modify or delete the paper archives that need to be modified, and record the time and operator of the modification or deletion; an archive backup module 140, which is used to regularly copy the paper archives in the electronic database to an external storage device; an archive statistics module 150, which is used to perform statistical analysis on the paper archives in the electronic database and generate charts or reports.

[0032] In the file entry module 110, the accuracy and completeness of the scan are ensured, and the assigned numbers and classification labels should be clear. By achieving a smooth transition from paper files to electronic databases, the use of paper files is reduced and the efficiency of file retrieval is improved. In the file query module 120, a variety of query conditions are provided, such as by medical record number, patient name, date range, etc., to ensure the accuracy and completeness of the query results. In this way, the speed and accuracy of file retrieval are improved, the retrieval time of staff is reduced, and work efficiency is improved. In the file modification module 130, the time, operator and specific operation content of each modification or deletion are recorded to ensure the traceability of the operation. In this way, the security and integrity of the file information are guaranteed, and convenience is provided for supervision and auditing. In the file backup module 140, backups are performed regularly, and the integrity and recoverability of the backup data are ensured. In this way, the security of the file information is improved and the data is guaranteed not to be lost due to accidents. In the file statistics module 150, a variety of statistical analysis functions are provided, such as statistical analysis by time, department, disease type, etc. By helping hospital management better understand the usage and trends of archival information, it provides data support for decision-making.

[0033] Among them, the text summary of paper documents can provide the main content and key information of paper archives, so that users can quickly understand the content of paper archives without browsing the entire archives. In addition, the text summary of paper documents can also be used as one of the bases for archive queries to improve the efficiency and accuracy of queries. However, in actual application scenarios, traditional archive management systems usually do not have the function of automatically generating text summaries, and manual extraction and writing are required. This method has the problems of low efficiency and long time consumption. Secondly, manual extraction and writing may result in uneven quality of text summaries due to subjective bias or omission of important information. Therefore, an optimized solution is expected.

[0034] Specifically, in the technical concept of the present application, text recognition technology and natural language processing technology based on deep learning are used to perform semantic analysis on paper archives, so as to intelligently generate text summaries of paper documents.

[0035] In one embodiment of the present application, the archive entry module 110 includes: a scanned file acquisition unit, used to acquire a scanned file of the paper archive to be analyzed; a text recognition unit, used to perform text recognition on the scanned file using text recognition technology to obtain a paper archive text description; a semantic feature extraction unit, used to extract semantic features of the paper archive text description to obtain a sequence of semantic encoding feature vectors of paper document text segment descriptions; an essential semantic information extraction unit, used to pass the sequence of semantic encoding feature vectors of paper document text segment descriptions through an essential semantic information extraction network to obtain a paper document text essential semantic information encoding feature vector; a text summary generation unit, used to obtain the paper document text summary based on the paper document text essential semantic information encoding feature vector.

[0036] The scanned file acquisition unit ensures that the quality and settings of the scanning device can obtain high-quality scanned files, avoid information loss or distortion, provide the basis for digital documents, and provide raw data for subsequent text recognition and information extraction. The text recognition unit uses efficient text recognition technology to ensure that the text in the scanned file can be accurately recognized and converted into electronic text, and converts the text information in the paper archive into a searchable and analyzable electronic text, which provides a basis for subsequent information extraction and analysis. The semantic feature extraction unit uses appropriate natural language processing technology to ensure that the extracted semantic features can accurately reflect the meaning and characteristics of the text, and converts the semantic information of the paper archive text into a computer-processable feature vector, which provides a basis for subsequent information extraction and text summary generation. The essential semantic information extraction unit ensures that the essential semantic information extraction network can accurately capture the essential meaning and important features of the text, and converts the essential semantic information of the text into a coded feature vector, which provides a basis for text summary generation and can help understand the core meaning of the text. The text summary generation unit uses a suitable text summary algorithm to ensure that the generated summary can accurately and concisely summarize the main information of the original text, generate a concise summary for the paper archive text, help users quickly understand the content of the document, and improve the readability and comprehensibility of the document.

[0037] Based on this, in the technical solution of the present application, the encoding process of the archive entry module 110 includes: first, obtaining a scanned file of the paper archive to be analyzed. Among them, the scanned file of the paper archive to be analyzed can be obtained using a scanning device such as a scanner. Then, text recognition is performed on the scanned file of the paper archive using text recognition technology to obtain a text description of the paper archive. Here, using text recognition technology to perform text recognition on the scanned file of the paper archive can extract text information from the image data to convert the content of the paper archive into text form. Through text recognition technology, text information can be automatically extracted from paper archives, avoiding the tedious process of manual input word by word and line by line.

[0038] In one embodiment of the present application, the semantic feature extraction unit includes: a data preprocessing subunit, used to perform data preprocessing on the paper archive text description to obtain a sequence of paper document text segment descriptions; a semantic coding analysis subunit, used to perform semantic coding analysis on the sequence of paper document text segment descriptions to obtain a sequence of semantic coding feature vectors of the paper document text segment descriptions.

[0039] Then, the paper archive text description is segmented to obtain a sequence of paper document text segment descriptions. Here, the continuous text description can be divided into paragraphs through segmentation, so as to divide the text description into a sequence of logically independent paper document text segment descriptions, so that the entire text structure is clearer. Furthermore, each paper document text segment description in the sequence of paper document text segment descriptions is passed through a text convolutional neural network model including a word embedding layer to obtain a sequence of semantic encoding feature vectors of paper document text segment descriptions. Among them, the word embedding layer can convert the words in the text segment description into vector representations, thereby providing rich semantic information. In the process of processing text data, the text convolutional neural network model can perform feature extraction and contextual semantic association learning on the vector representation of each word to capture its implicit semantic feature information.

[0040] In a specific embodiment of the present application, the data preprocessing subunit is used to: segment the paper archive text description to obtain a sequence of paper document text segment descriptions.

[0041] Furthermore, in a specific embodiment of the present application, the semantic coding analysis subunit is used to: pass each paper document text segment description in the sequence of the paper document text segment description through a text convolutional neural network model including a word embedding layer to obtain a sequence of semantic coding feature vectors of the paper document text segment description.

[0042] Among them, the semantic coding analysis subunit includes: a secondary subunit of word segmentation, which is used to perform word segmentation processing on each paper document text segment description in the sequence of paper document text segment descriptions to convert each paper document text segment description in the sequence of paper document text segment descriptions into a word sequence composed of multiple words; a secondary subunit of embedded coding, which is used to use the word embedding layer of the text convolutional neural network model including the word embedding layer to map each word in the word sequence to a word vector to obtain a sequence of word vectors; a secondary subunit of context coding, which is used to use the text convolutional neural network model including the word embedding layer to semantically encode the sequence of word vectors to obtain a sequence of semantic coding feature vectors of the paper document text segment descriptions.

[0043] Considering that people may use repetitive, lengthy or unnecessary expressions in the process of writing and recording information. These redundant information may be caused by factors such as the author's expression, writing habits, and language style. In any case, these redundant information exist in paper documents to a greater or lesser extent and are retained in the process of extracting semantic feature information, which will affect the accuracy of generating paper document text summaries. In order to reduce the impact of these redundant information on paper document text summaries, in the technical solution of the present application, the sequence of semantic encoding feature vectors describing the paper document text segment is passed through an essential semantic information extraction network to capture the core meaning and semantic expression of the text, thereby obtaining a paper document text essential semantic information encoding feature vector. Specifically, the essential semantic information extraction network can automatically select and emphasize features that have a greater contribution to the task of generating paper document text summaries, while weakening or ignoring those redundant features. This can reduce the impact of redundant information and noise in paper documents and retain more important and useful semantic information.

[0044] In a specific embodiment of the present application, the essential semantic information extraction unit is used to: process the sequence of semantic encoding feature vectors describing the paper document text segment using the following essential semantic information extraction formula to obtain the paper document text essential semantic information encoding feature vector; wherein the essential semantic information extraction formula is: ; ; in, is the first in the sequence of semantic coding feature vectors describing the paper document text segment A paper document text segment describes the semantic encoding feature vector, is the first in the sequence of semantic coding feature vectors describing the paper document text segment A paper document text segment describes the semantic encoding feature vector, represents the 1-norm of the eigenvector, The length of the sequence of semantic encoding feature vectors describing the paper document text segment - 1, A representation of a sequence of semantic encoding feature vectors describing the paper document text segment, represents the semantic feature difference coefficient, represents the natural exponential function operation, represents the total number of difference coefficients of the semantic features, A feature vector is encoded for the essential semantic information of the paper document text.

[0045] Furthermore, the paper document text essential semantic information encoding feature vector is passed through an AIGC-based archive summary generator to obtain a paper document text summary. In this way, the AIGC (Attention-basedInformationGatedConvolution)-based archive summary generator is used to generate a text summary of a paper document to provide a summary and key information of the document content, which is used to quickly understand the document content and retrieve the document.

[0046] In a specific embodiment of the present application, the text summary generation unit includes: an optimization subunit, used to optimize the paper document text essential semantic information encoding feature vector to obtain an optimized paper document text essential semantic information encoding feature vector; an archive summary generation subunit, used to pass the optimized paper document text essential semantic information encoding feature vector through an AIGC-based archive summary generator to obtain the paper document text summary.

[0047] In the technical solution of the present application, each paper document text segment description semantic coding feature vector in the sequence of paper document text segment description semantic coding feature vectors expresses the text coding semantic features of the corresponding segment of the paper archive text description, so that after the sequence of paper document text segment description semantic coding feature vectors passes through the essential semantic information extraction network, it can be weighted based on the essential importance of the semantic feature distribution of each paper document text segment description semantic coding feature vector relative to the overall semantic feature distribution, so that the paper document text segment description semantic coding feature vector has a differential expression based on its semantic essentiality. However, this will also cause the feature representation of the paper document text essential semantic information coding feature vector to deviate from the text coding semantic feature representation of the sequence of paper document text segment description semantic coding feature vectors, so that the paper document text summary obtained by the paper document text essential semantic information coding feature vector through the archive summary generator based on AIGC is less consistent with the source text description.

[0048] Based on this, it is considered to further fuse the feature vector encoding the essential semantic information of the paper document text with the sequence of the semantic encoding feature vector of the paper document text segment description to overcome the above-mentioned defects. In addition, considering the difference in feature distribution information representation between the feature vector encoding the essential semantic information of the paper document text and the sequence of the semantic encoding feature vector of the paper document text segment description due to essential semantic weighting, it will cause the semantic aggregation distribution richness of the fused feature vector encoding the essential semantic information of the paper document text, resulting in repetition of unstructured probability mapping, affecting the accuracy of the generated results.

[0049] Therefore, preferably, when the paper document text summary is obtained by passing the paper document text essential semantic information encoding feature vector through the archive summary generator based on AIGC, the paper document text essential semantic information encoding feature vector is optimized. Specifically, the optimization includes the following steps: The length of the feature vector encoding the essential semantic information of the paper document text Subtracting the zero norm of the feature vector encoding the essential semantic information of the paper document text to determine the number of zero eigenvalues ​​of the feature vector encoding the essential semantic information of the paper document text , and the number of zero eigenvalues Subtract one to get the zero differentiable value of the essential semantic information encoding of the paper document text ; Encode the essential semantic information of the paper document text into a zero differentiable value Multiply and divide by the number of zero eigenvalues ​​respectively Get the first paper document text essential semantic information encoding field fitting value and the fitting value of the encoding field of the essential semantic information of the second paper document text ; Calculate the square root of the sum of squares of all eigenvalues ​​of the paper document text essential semantic information encoding feature vector to obtain the paper document text essential semantic information encoding mode representation value ,in, The first character representing the encoding feature vector of the essential semantic information of the paper document text eigenvalues, The total number of eigenvalues ​​of the feature vector encoding the essential semantic information of the paper document text, The encoding mode representation value representing the essential semantic information of the paper document text; Encoding zero differentiable values ​​with the essential semantic information of the paper document text The power function of each eigenvalue of the feature vector encoding the essential semantic information of the paper document text is calculated as an exponent , and multiply the fitting value of the encoding field of the essential semantic information of the first paper document text by To obtain the first paper document text essential semantic information encoding intermediate representation vector ; The paper document text essential semantic information encoding feature vector and the second paper document text essential semantic information encoding field fitting value are combined. and the paper document text essential semantic information encoding mode representation value Perform point multiplication to obtain the second paper document text essential semantic information encoding intermediate representation vector ; Calculate the intermediate representation vector encoding the essential semantic information of the first paper document text The intermediate representation vector encoding the essential semantic information of the second paper document text The weighted points minus To obtain the optimized paper document text essential semantic information encoding feature vector, where: and represents the weighted hyperparameter, It means point multiplication by position. Indicates point subtraction by position.

[0050] Finally, the optimized paper document text essential semantic information encoding feature vector is passed through an AIGC-based archive summary generator to obtain the paper document text summary.

[0051] That is, based on the vector field fitting law of high-dimensional manifold, the isolated zeros in the feature vector encoding the essential semantic information of the paper document text are used as the differentiable dimension of the high-dimensional manifold, so as to fix the normal direction pointing to the distribution center near the local coordinates represented by the eigenvalues ​​of the feature vector encoding the essential semantic information of the paper document text, thereby avoiding invalid repetition in the probability mapping process caused by the unstructured characteristics of the rich distribution of the feature vector encoding the essential semantic information of the paper document text by effectively aligning the characteristic distribution pattern of the feature vector encoding the essential semantic information of the paper document text to the probability density differentiable field, thereby improving the text quality of the paper document text summary obtained by the archive summary generator based on AIGC.

[0052] In summary, the archive management system 100 based on the embodiment of the present application is explained, which uses text recognition technology and natural language processing technology based on deep learning to perform semantic analysis on paper archives, thereby intelligently generating a text summary of the paper document.

[0053] As described above, the archive management system 100 according to the embodiment of the present application can be implemented in various terminal devices, such as a server for archive management. In one example, the archive management system 100 according to the embodiment of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the archive management system 100 can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the archive management system 100 can also be one of the many hardware modules of the terminal device.

[0054] Alternatively, in another example, the archive management system 100 and the terminal device may also be separate devices, and the archive management system 100 may be connected to the terminal device via a wired and / or wireless network, and transmit interactive information in accordance with an agreed data format.

[0055] Figure 2 Flow chart of a file management method provided in an embodiment of the present application. Figure 2 As shown, a file management method includes: 210, scanning paper files into an electronic database, assigning a number and classification label to each of the paper files, and generating a paper document text summary for each of the paper files; 220, searching for the required paper files according to the query conditions, and displaying the detailed information of the required paper files; 230, modifying or deleting the paper files that need to be modified, and recording the time and operator of the modification or deletion; 240, regularly copying the paper files in the electronic database to an external storage device; 250, performing statistical analysis on the paper files in the electronic database and generating charts or reports.

[0056] Figure 3 Schematic diagram of the system architecture of a file management method provided in an embodiment of the present application. Figure 3 As shown, paper archives are scanned into an electronic database, a number and a classification label are assigned to each of the paper archives, and a paper document text summary is generated for each of the paper archives, including: first, obtaining a scanned file of the paper archive to be analyzed; then, using text recognition technology to perform text recognition on the scanned file to obtain a paper archive text description; then, extracting the semantic features of the paper archive text description to obtain a sequence of semantic encoding feature vectors of paper document text segment descriptions; then, passing the sequence of semantic encoding feature vectors of paper document text segment descriptions through an essential semantic information extraction network to obtain a paper document text essential semantic information encoding feature vector; finally, based on the paper document text essential semantic information encoding feature vector, the paper document text summary is obtained.

[0057] In the archive management method, paper archives are scanned into an electronic database, a number and a classification label are assigned to each of the paper archives, and a paper document text summary is generated for each of the paper archives, including: obtaining a scanned file of the paper archive to be analyzed; using text recognition technology to perform text recognition on the scanned file to obtain a paper archive text description; extracting semantic features of the paper archive text description to obtain a sequence of semantic encoding feature vectors of paper document text segment descriptions; passing the sequence of semantic encoding feature vectors of paper document text segment descriptions through an essential semantic information extraction network to obtain a paper document text essential semantic information encoding feature vector; and obtaining the paper document text summary based on the paper document text essential semantic information encoding feature vector.

[0058] Those skilled in the art will appreciate that the specific operations of each step in the above-mentioned archive management method have been described in detail above. Figure 1 The description of the archive management system has been described in detail, and therefore, its repeated description will be omitted.

[0059] Figure 4 This is an application scenario diagram of an archive management system provided in an embodiment of the present application. Figure 4 As shown, in this application scenario, first, a scanned file of the paper archive to be analyzed is obtained (for example, Figure 4 Then, the scanned files of the acquired paper archives are input into a server deployed with an archive management algorithm (for example, Figure 4 In S) shown in , the server is capable of processing the scanned file of the paper archive based on the archive management algorithm to obtain the text summary of the paper document.

[0060] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the scope of protection of the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A file management system, characterized in that: include: The file entry module is used to scan paper files into the electronic database, assign a number and classification label to each paper file, and generate a paper document text summary for each paper file; The file query module is used to search for the required paper files according to the query conditions and display the detailed information of the required paper files; The file modification module is used to modify or delete the paper files that need to be modified, and record the time of modification or deletion and the operator; An archive backup module, used to regularly copy paper archives in the electronic database to an external storage device; An archive statistics module is used to perform statistical analysis on paper archives in the electronic database and generate charts or reports; The file entry module includes: A scanned file acquisition unit, used to acquire the scanned file of the paper archive to be analyzed; A text recognition unit, used for performing text recognition on the scanned document using text recognition technology to obtain a paper archive text description; A semantic feature extraction unit, used to extract the semantic features of the paper archive text description to obtain a sequence of semantic coding feature vectors describing the paper document text segment; An essential semantic information extraction unit is used to pass the sequence of semantic coding feature vectors describing the paper document text segment through an essential semantic information extraction network to obtain an essential semantic information coding feature vector of the paper document text; The text summary generating unit is used to encode a feature vector based on the essential semantic information of the paper document text to obtain the paper document text summary.

2. The file management system according to claim 1, characterized in that: The semantic feature extraction unit comprises: A data preprocessing subunit, used for performing data preprocessing on the paper archive text description to obtain a sequence of paper document text segment descriptions; The semantic coding analysis subunit is used to perform semantic coding analysis on the sequence of the paper document text segment description to obtain a sequence of semantic coding feature vectors of the paper document text segment description.

3. The file management system according to claim 2, characterized in that: The data preprocessing subunit is used to: The paper archive text description is segmented to obtain a sequence of paper document text segment descriptions.

4. The file management system according to claim 3, characterized in that: The semantic coding analysis subunit is used to: Each paper document text segment description in the sequence of the paper document text segment description is passed through a text convolutional neural network model including a word embedding layer to obtain a sequence of semantic encoding feature vectors of the paper document text segment description.

5. The file management system according to claim 4, characterized in that: The semantic coding analysis subunit comprises: A word segmentation secondary subunit is used to perform word segmentation processing on each paper document text segment description in the sequence of paper document text segment descriptions to convert each paper document text segment description in the sequence of paper document text segment descriptions into a word sequence composed of multiple words; An embedding coding secondary subunit, configured to map each word in the word sequence to a word vector using the word embedding layer of the text convolutional neural network model including the word embedding layer to obtain a sequence of word vectors; The context encoding secondary subunit is used to use the text convolutional neural network model including the word embedding layer to semantically encode the sequence of word vectors to obtain a sequence of semantic encoding feature vectors describing the paper document text segment.

6. The file management system according to claim 5, characterized in that: The essential semantic information extraction unit is used to: The sequence of semantic coding feature vectors describing the paper document text segment is processed using the following essential semantic information extraction formula to obtain the paper document text essential semantic information coding feature vector; wherein the essential semantic information extraction formula is: ； ； in, is the first in the sequence of semantic coding feature vectors describing the paper document text segment A paper document text segment describes the semantic encoding feature vector, is the first in the sequence of semantic coding feature vectors describing the paper document text segment A paper document text segment describes the semantic encoding feature vector, represents the 1-norm of the eigenvector, The length of the sequence of semantic encoding feature vectors describing the paper document text segment - 1, A representation of a sequence of semantic encoding feature vectors describing the paper document text segment, represents the semantic feature difference coefficient, represents the natural exponential function operation, represents the total number of difference coefficients of the semantic features, A feature vector is encoded for the essential semantic information of the paper document text.

7. The file management system according to claim 6, characterized in that: The text summary generating unit comprises: An optimization subunit, used for optimizing the paper document text essential semantic information encoding feature vector to obtain an optimized paper document text essential semantic information encoding feature vector; The archive summary generation subunit is used to pass the optimized paper document text essential semantic information encoding feature vector through an AIGC-based archive summary generator to obtain the paper document text summary.

8. A file management method, characterized in that: include: Scanning paper archives into an electronic database, assigning a number and classification label to each paper archive, and generating a paper document text summary for each paper archive; According to the query conditions, the required paper archives are searched and the detailed information of the required paper archives is displayed; Modify or delete paper files that need to be modified, and record the time of modification or deletion and the operator; Periodically copy the paper records in the electronic database to an external storage device; Conduct statistical analysis on the paper files in the electronic database and generate charts or reports; The paper archives are scanned into an electronic database, a number and a classification label are assigned to each of the paper archives, and a paper document text summary is generated for each of the paper archives, including: Obtaining a scanned file of the paper archive to be analyzed; Using text recognition technology to perform text recognition on the scanned document to obtain a paper archive text description; Extracting semantic features of the paper archive text description to obtain a sequence of semantic encoding feature vectors describing the paper document text segment; Passing the sequence of semantic coding feature vectors describing the paper document text segment through an essential semantic information extraction network to obtain a paper document text essential semantic information coding feature vector; Based on the feature vector encoded with the essential semantic information of the paper document text, a summary of the paper document text is obtained.

Images