A software requirement document analysis method, device and medium

Abstract

The application discloses a software requirement document analysis method and device and medium, and belongs to the technical field of document analysis, and is used for solving the technical problems that the information analysis in the existing requirement document has content deviation, manual analysis and extraction are complex and tedious, and the requirement document cannot be efficiently and accurately analyzed. The method comprises the following steps: performing layout analysis on text information and image information of the requirement document to obtain layout area information; classifying and positioning a plurality of elements in the page requirement information to obtain first key information based on the page requirement information; performing text analysis on the word sequence of the function requirement information to obtain second key information based on the function requirement information; performing layout recognition training on the requirement document based on the first key information and the second key information to obtain a comprehensive requirement model; and defining the requirement document to be analyzed to generate corresponding software development parameter information.

Description

Technical Field

[0001] The present application relates to the field of document parsing, and in particular to a method, device, and medium for parsing software requirement documents. Background Art

[0002] Requirements documents play a critical role in the software development process, providing the development team with a foundation and shared understanding of the project. However, as software projects grow in size and complexity, parsing requirements documents becomes increasingly challenging. Manually parsing requirements documents is not only time-consuming but also prone to errors and misunderstandings, potentially leading to confusion and misinterpretation during the development process.

[0003] Parsing the interface requirements and functional requirements in requirements documents is a crucial step in the software development process. Interface requirements describe the appearance, layout, and interaction of the user interface, while functional requirements specify the functionality and behavior that the system must implement. However, traditional parsing methods often require manual segmentation and extraction of these requirements, a tedious and error-prone process in large-scale projects. Summary of the Invention

[0004] The embodiments of the present application provide a software requirement document parsing method, device and medium for solving the following technical problems: the information parsing in existing requirement documents has content deviations, and manual parsing and extraction are complex and tedious, making it difficult to parse the requirement documents efficiently and accurately.

[0005] The embodiments of this application adopt the following technical solutions:

[0006] On the one hand, an embodiment of the present application provides a method for parsing a software requirement document, comprising: performing layout analysis on the text information and image information of the requirement document to obtain layout area information; wherein the layout area information includes: page requirement information and function requirement information; classifying and locating multiple elements in the page requirement information through a preset convolutional neural network to obtain first key information based on the page requirement information; performing text parsing on the word sequence of the function requirement information to obtain second key information based on the function requirement information; performing layout recognition training on the requirement document based on the first key information and the second key information to obtain a comprehensive requirement model; and defining the requirements of the requirement document to be parsed through the comprehensive requirement model to generate corresponding software development parameter information.

[0007] The embodiment of the application realizes intelligent analysis of the requirement document by combining layout analysis technology, computer vision technology and natural language processing technology. Meanwhile, the embodiment has the characteristics of high efficiency, accuracy and automation, and can greatly improve the analysis efficiency and reduce the generation of human errors. By generating the page requirement model and the function requirement model, and then combining the page requirement model and the function requirement model into the comprehensive requirement model, the requirement definition of the requirement document to be analyzed is obtained, and clear requirement definition and design guidance are provided for the software development team, which helps to improve the quality and efficiency of software development.

[0008] In a feasible implementation, the layout analysis on the text information and the image information of the requirement document is performed to obtain the layout area information, specifically including: performing full-text scanning on the requirement document by using a preset image processing algorithm to obtain a full-text layout image; performing feature recognition on the full-text layout image based on an image feature recognition box, and uniformly marking the recognized image features to obtain image marked features; performing image area recognition and marking on the text features in the full-text layout image by using a preset OCR recognition technology to obtain a first marked area; performing area division on the full-text layout image according to the main features of the requirement document, and marking a second marked area; wherein the main features include document layout features, document title features and document key marker features of the requirement document; the document content attributes include text attributes, title attributes, table attributes and picture attributes of the requirement document; based on the first marked area and the second marked area, the image marked features are classified according to the layout area division to determine the layout area information.

[0009] In a feasible implementation, based on the first marked area and the second marked area, the image marked features are classified according to the layout area division to determine the layout area information, specifically including: performing image area feature matching on the first marked area and the image marked area, and determining a region meeting the feature matching as a function requirement region; performing image area feature matching on the second marked area and the image marked area, and determining a region meeting the feature matching as a page requirement region; performing layout area division on the full-text layout image according to the function requirement region and the page requirement region to obtain the layout area information.

[0010] In an implementable embodiment, the plurality of elements in the page requirement information are classified and located by a preset convolutional neural network to obtain first key information based on the page requirement information, specifically including: based on a preset image target detection technology, performing feature recognition on the page requirement information about a plurality of element attributes to obtain element features; wherein the element features include: page template type features, icon type features, text embedding features, and table style features; performing feature positioning on the plurality of elements in the page requirement information after recognition by the convolutional neural network to obtain positioning information; based on the positioning information and the element features, performing same-category element classification learning training on the plurality of elements to obtain a deep learning convolutional neural network; based on the deep learning convolutional neural network, positioning and classifying each element of the to-be-recognized page requirement information to determine the first key information based on the page requirement information.

[0011] In an implementable embodiment, after positioning and classifying each element of the to-be-recognized page requirement information based on the deep learning convolutional neural network, specifically including: by an OCR recognition technology, performing text conversion on the text embedding features in the to-be-recognized page requirement information to obtain editable text features; based on the deep learning convolutional neural network, positioning the text features, and determining the positioned text features as the first key information; wherein the first key information includes the positioned text feature information and the positioning and classification information of each element.

[0012] In an implementable embodiment, the function requirement information is subjected to text analysis on word sequences to obtain second key information based on the function requirement information, specifically including: by a preset natural language processing technology, performing word division on continuous text in the function requirement information to obtain word sequences; performing text word type identification and classification on the word sequences to obtain word type information; wherein the word type information includes: management data information, query data information, and application data information; based on a named entity recognition technology, performing text entity recognition on the word type information to obtain specific entity information; wherein the specific entity information includes: date information, name information, location information, and requirement explanation information; performing dependency syntax analysis on the word sequences to determine the syntax structure and dependency relationship information of the function requirement information; based on the word sequences, the word type information, the specific entity information, and the syntax structure and dependency relationship information, extracting key information from the function requirement information to obtain the second key information; wherein the second key information includes: field name information, function type information, field length information, business rule information, and dependency relationship information.

[0013] In a possible implementation, the requirement document is subjected to layout recognition training based on the first key information and the second key information to obtain a comprehensive requirement model, specifically including: the requirement document is subjected to generation training about a page template based on the first key information to obtain a page requirement model; the requirement document is subjected to generation training about a function template based on the first key information to obtain a function requirement model; the page requirement model and the function requirement model are subjected to the same association processing about the requirement document to determine the comprehensive requirement model.

[0014] In a possible implementation, the page template includes a page type element, an icon type element, a character text feature element, and a table style element; and the function requirement model includes a word order element, a word type element, a specific entity element, a syntax structure element, and a dependency relationship element.

[0015] In a second aspect, the embodiments of the present application further provide a software requirement document analysis device, the device comprising: at least one processor; and a memory in communication connection with the at least one processor; wherein the memory stores instructions executable by the at least one processor, so that the at least one processor can execute the software requirement document analysis method of any of the above-mentioned embodiments.

[0016] In a third aspect, the embodiments of the present application further provide a non-volatile computer storage medium, characterized in that the storage medium is a non-volatile computer readable storage medium, and the non-volatile computer readable storage medium stores at least one program, each of the programs including instructions, which, when executed by a terminal, cause the terminal to execute the software requirement document analysis method of any of the above-mentioned embodiments.

[0017] The present application provides a software requirement document analysis method, device and medium, compared with the prior art, the embodiments of the present application have the following beneficial technical effects:

[0018] The embodiments of the present application realize intelligent analysis of requirement documents by combining layout analysis technology, computer vision technology and natural language processing technology. Meanwhile, the embodiments have the characteristics of high efficiency, accuracy and automation, and can greatly improve the analysis efficiency and reduce the generation of human errors. Through the generation of the page requirement model and the function requirement model, the page requirement model and the function requirement model are combined into a comprehensive requirement model, and then the requirement definition of the requirement document to be analyzed is obtained, which provides clear requirement definition and design guidance for software development teams, and helps to improve the quality and efficiency of software development. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to make the technical solutions in the embodiments of the present application or the prior art clearer, the accompanying drawings needed in the embodiments or prior art description will be briefly introduced. Obviously, the accompanying drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort. In the drawings:

[0020] Figure 1 A software requirement document analysis method flow chart provided by the embodiments of the present application;

[0021] Figure 2 A software requirement document analysis method block diagram provided by the embodiments of the present application;

[0022] Figure 3 A page requirement analysis effect related to icon recognition provided by the embodiments of the present application;

[0023] Figure 4 A structural schematic diagram of a software requirement document analysis device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0024] In order for those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the accompanying drawings of the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without any creative effort should be within the scope of protection of the present application.

[0025] The embodiments of the present application provide a software requirement document analysis method, as shown in Figure 1 The software requirement document analysis method specifically includes steps S101-S105:

[0026] It should be noted that in software development, the requirement document plays a bridging role, providing the development team with a common understanding of the project goals, functions and constraints. However, the requirement document often has a complex layout structure and a large amount of information, and manual analysis is time-consuming and prone to errors. Therefore, the goal of the present application is to provide an automated analysis method for requirement documents to improve the efficiency and accuracy of requirement document analysis.

[0027] S101, the requirement document is analyzed for text information and image information, and layout area information is obtained. The layout area information includes page requirement information and function requirement information.

[0028] Specifically, first, the demand document is full-text scanned by a preset image processing algorithm to obtain a full-text layout image. Then, image features are recognized based on an image feature recognition box, and the recognized image features are uniformly marked to obtain image marked features.

[0029] Further, text features in the full-text layout image are recognized and marked by a preset OCR recognition technology to obtain a first marked region. According to the main features of the demand document, the full-text layout image is regionally divided in relation to document content attributes, and a second marked region is marked. The main features include document layout features, document title features, and document key marker features of the demand document. The document content attributes include text attributes, title attributes, table attributes, and picture attributes of the demand document.

[0030] Further, the image marked features are regionally divided and classified in relation to the layout based on the first marked region and the second marked region to determine layout region information.

[0031] The first marked region and the image marked region are first regionally matched in image features, and the region that meets the feature matching is determined as a functional requirement region. Then, the second marked region and the image marked region are regionally matched in image features, and the region that meets the feature matching is determined as a page requirement region. Finally, the full-text layout image is regionally divided in relation to the layout based on the functional requirement region and the page requirement region to obtain layout region information.

[0032] In one embodiment, Figure 2 A software requirement document analysis method diagram provided by the embodiment of the present application is shown in FIG. 1. Figure 2 First, the demand document is intelligently divided by using layout analysis technology. Layout analysis is a process of dividing a text image into different layout regions. By combining deep learning and image processing algorithms, the layout features, title features, and marker features of the full-text layout image in the demand document are recognized. Then, the text features in the full-text layout image are regionally recognized and marked, and the full-text layout image is regionally divided and marked in relation to document content attributes. Based on the first marked region and the second marked region, the system divides the demand document into ten categories of regions, such as text, title, table, and picture. The text in each region is extracted by OCR recognition. The region that meets the feature matching is determined as a functional requirement region and a page requirement region, and the page requirement information and the functional requirement information are automatically separated and determined. This automatic layout region division not only saves the time and labor cost of manual division, but also reduces the problems caused by misunderstanding or incorrect division.

[0033] S102, classify and locate the plurality of elements in the page requirement information by the preset convolutional neural network, to obtain first key information based on the page requirement information.

[0034] Specifically, based on the preset image target detection technology, the page requirement information is subjected to feature recognition on the attributes of the plurality of elements, to obtain element features. The element features include: page template type features, icon type features, text embedding features, and table style features.

[0035] Further, the plurality of elements in the recognized page requirement information are subjected to feature positioning by the convolutional neural network, to obtain positioning information.

[0036] Further, according to the positioning information and the element features, the plurality of elements are subjected to classification learning and training of elements of the same category, to obtain the convolutional neural network after deep learning.

[0037] Further, based on the convolutional neural network after deep learning, the elements in the page requirement information to be recognized are subjected to positioning and classification, to determine the first key information based on the page requirement information.

[0038] Further, based on the convolutional neural network after deep learning, the elements in the page requirement information to be recognized are subjected to positioning and classification, to determine the first key information based on the page requirement information.

[0039] In one embodiment, as shown in Figure 2 For the page requirement information part, computer vision technology, target detection technology and OCR technology are adopted. Image classification and target detection technology can recognize the element features in the page design diagram, such as page type template features, icon features, text features and table style features, etc. Figure 3 The page requirement analysis effect provided by the embodiment of the present application is related to icon recognition, as shown in Figure 3The matching degree of each icon identified in the figure is shown, i.e., in the figure. Then, by applying a deep learning model and a convolutional neural network, a plurality of elements in the identified page requirement information are subjected to feature positioning, and then based on the element features, the plurality of elements are subjected to classification learning training of elements of the same category to obtain a deep learning convolutional neural network, so as to realize positioning and classification of each element of the to-be-identified page requirement information, and determine the first key information. At the same time, the OCR technology can also convert the text in the to-be-identified page requirement information into editable text, and further extract the key information of the page requirement information.

[0040] In S103, text analysis on the sequence of related words of the function requirement information is performed to obtain second key information based on the function requirement information.

[0041] Specifically, by using a preset natural language processing technology, continuous text in the function requirement information is subjected to word division to obtain a word sequence.

[0042] Further, the word sequence is subjected to identification and classification of text word types to obtain word type information. The word type information includes: management data information, query data information, and application data information.

[0043] Further, based on a named entity recognition technology, the word type information is subjected to text entity recognition to obtain specific entity information. The specific entity information includes: date information, name information, location information, and requirement explanation information.

[0044] At the same time, the word sequence is subjected to dependency syntax analysis to determine the syntax structure and dependency relationship information of the function requirement information.

[0045] Further, based on the word sequence, the word type information, the specific entity information, and the syntax structure and dependency relationship information, the function requirement information is subjected to extraction of key information to obtain the second key information. The second key information includes: field name information, function type information, field length information, business rule information, and dependency relationship information.

[0046] In one embodiment, as shown in FIG. 1, the system for extracting key information of a page requirement information includes a page requirement information obtaining unit 101, a function requirement information obtaining unit 102, a function requirement information processing unit 103, and a key information extraction unit 104. Figure 2As shown, for the functional requirement information part, natural language processing techniques are used for text analysis. The word segmentation technique can divide continuous text into a sequence of words, providing a basis for subsequent text processing and analysis. The text classification technique can identify the type of functional requirement text, such as user management, data query, etc. The named entity recognition technique can identify specific entities in the text, such as dates, names, places, etc. The dependency syntax analysis technique can analyze the syntax structure and dependency relationship in the sentence. By combining these natural language processing techniques, the field name, type, length, business rules, dependency relationship, and other key information can be extracted from the text of the functional requirement information. Then, based on the word sequence, word type information, specific entity information, and syntax structure and dependency relationship information, the key information of the functional requirement information is extracted, thereby obtaining the second key information based on the functional information part.

[0047] S104, based on the first key information and the second key information, performing layout recognition training on the requirement document to obtain a comprehensive requirement model.

[0048] Specifically, based on the first key information, the requirement document is trained for generating a page template to obtain a page requirement model. At the same time, based on the first key information, the requirement document is trained for generating a functional template to obtain a functional requirement model. Finally, the page requirement model and the functional requirement model are associated with the same requirement document to determine the comprehensive requirement model.

[0049] The page template includes a page type element, an icon type element, a text feature element, and a table style element. The functional requirement model includes a word sequence element, a word type element, a specific entity element, a syntax structure element, and a dependency relationship element.

[0050] In one embodiment, by identifying and positioning training of each element in the page template and the functional requirement model, the corresponding page requirement model and the functional requirement model are generated. The combination of the page requirement model and the functional requirement model, i.e., the comprehensive requirement model, can provide a reference for subsequent page development and design, and also provide accurate requirement definition and analysis for subsequent software development and other work.

[0051] S105, by the comprehensive requirement model, performing requirement definition on the requirement document to be analyzed to generate corresponding software development parameter information.

[0052] Specifically, based on the comprehensive demand model, automatic parsing and information extraction of the demand document are realized, thereby improving the efficiency and accuracy of parsing. Meanwhile, by combining image classification, target detection, OCR, word segmentation, text classification and other technologies, better tools and methods are provided for the software development team, and corresponding software development parameter information is generated, which is more conducive to the automatic parsing of the recognized demand document.

[0053] In one embodiment, the embodiments of the present application also have:

[0054] 1. Efficiency: The demand document is parsed in an automated manner, saving a lot of manual partitioning and information extraction time, and greatly improving the parsing efficiency.

[0055] 2. Accuracy: With the aid of computer vision and natural language processing technology, the system can accurately extract key information from the page demand design diagram and functional demand text, avoiding misunderstandings and errors in the human parsing process.

[0056] 3. Automation: The method of the present application can automatically divide page demand and functional demand, and generate the corresponding demand model, reducing the demand modeling process of manual intervention and improving the degree of automation.

[0057] In addition, the embodiments of the present application also provide a software demand document parsing device, as shown in Figure 4 The software demand document parsing device 400 specifically includes:

[0058] At least one processor 401; and a memory 402 communicatively connected with the at least one processor 401; wherein the memory 402 stores instructions executable by the at least one processor 401, so that the at least one processor 401 can execute:

[0059] Performing layout analysis on the text information and image information of the demand document to obtain layout area information; wherein the layout area information includes: page demand information and functional demand information;

[0060] Classifying and positioning a plurality of elements in the page demand information through a preset convolutional neural network to obtain first key information based on the page demand information;

[0061] Performing text analysis on the word sequence of the functional demand information to obtain second key information based on the functional demand information;

[0062] Performing layout recognition training on the demand document based on the first key information and the second key information to obtain a comprehensive demand model;

[0063] Defining the demand of the demand document to be parsed through the comprehensive demand model to generate corresponding software development parameter information.

[0064] The application provides a software requirement document analysis method, device and medium, which realizes intelligent analysis of the requirement document by combining layout analysis technology, computer vision technology and natural language processing technology. Meanwhile, the application has the characteristics of high efficiency, accuracy and automation, and can greatly improve the analysis efficiency and reduce the generation of human errors. By generating a page requirement model and a function requirement model, and then combining the page requirement model and the function requirement model into a comprehensive requirement model, the requirement definition of the requirement document to be analyzed is obtained, and clear requirement definition and design guidance are provided for a software development team, which helps to improve the quality and efficiency of software development.

[0065] Each embodiment in the application is described in a progressive manner, and the same or similar parts of each embodiment can be referred to each other. Each embodiment mainly describes the difference from other embodiments. Especially, the device and non-volatile computer storage medium embodiment is basically similar to the method embodiment, so the description is relatively simple, and the related parts can be referred to the part of the method embodiment.

[0066] The device and medium provided by the embodiments of the application are one-to-one corresponding to the method, so the device and medium also have similar beneficial technical effects to the method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the device and medium will not be described here.

[0067] Those skilled in the art should understand that the embodiments of the application can be provided as a method, a system, or a computer program product. Therefore, the application can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the application can be in the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0068] The application is described with reference to flowcharts and / or block diagrams of the method, device (system), and computer program product according to the embodiments of the application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a machine that implements the function specified in the flowcharts and / or block diagrams. Figure 1 The function specified in one flow or multiple flows and / or blocks Figure 1 The device that implements the function specified in one block or multiple blocks.

[0069] Memory can include non-persistent memory, Random Access Memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash memory, in computer readable media. Memory is an example of computer readable media.

[0070] Computer readable media includes permanent and non-permanent, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile discs (DVDs) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer readable media does not include transitory media, such as modulated data signals and carrier waves.

[0071] The above-described embodiments of the application have been described in connection with what are presently considered to be the most practical and preferred implementations. From the foregoing description, it will be apparent to those skilled in the art that various modifications and variations can be made to the present application without departing from the spirit and scope of the application. It is, therefore, intended that this application be interpreted as including all such modifications and variations as can fairly be said to be equivalent to the described embodiments. It is intended that the application embrace all alternatives fairly within the scope of the description.

[0072] The above description is intended to be illustrative and not restrictive. Many embodiments of the present application will be apparent to those of skill in the art upon reviewing the above description. The scope of the application should, therefore, be determined not with reference to the above description, but instead with reference to the appended claims, along with their full scope of equivalents.

Claims

1. A software requirement document parsing method, characterized by, The method comprises: performing layout analysis on the demand document regarding text information and image information to obtain layout area information; wherein the layout area information comprises page demand information and function demand information; performing classification and positioning on multiple elements in the page demand information through a preset convolutional neural network to obtain first key information based on the page demand information, specifically comprising: performing feature recognition on multiple element attributes of the page demand information based on a preset image target detection technology to obtain element features; wherein the element features comprise page template type features, icon type features, text embedding features, and table style features; performing feature positioning on multiple elements in the page demand information after feature recognition through the convolutional neural network to obtain positioning information; performing classification learning training on multiple elements of the same category according to the positioning information and through the element features to obtain a convolutional neural network after deep learning; performing positioning and classification on each element of the page demand information to be recognized based on the convolutional neural network after deep learning to determine the first key information based on the page demand information; performing text analysis on the function demand information regarding word sequences to obtain second key information based on the function demand information, specifically comprising: performing word division on continuous text in the function demand information through a preset natural language processing technology to obtain word sequences; performing identification and classification of text word types on the word sequences to obtain word type information; wherein the word type information comprises management data information, query data information, and application data information; performing text entity recognition on the word type information based on a named entity recognition technology to obtain specific entity information; wherein the specific entity information comprises date information, name information, location information, and demand explanation information; performing dependency syntax analysis on the word sequences to determine the syntax structure and dependency relationship information of the function demand information; performing key information extraction on the function demand information based on the word sequences, the word type information, the specific entity information, and the syntax structure and dependency relationship information to obtain the second key information; wherein the second key information comprises field name information, function type information, field length information, business rule information, and dependency relationship information; performing layout recognition training on the demand document based on the first key information and the second key information to obtain a comprehensive demand model; defining demand of a demand document to be analyzed through the comprehensive demand model to generate corresponding software development parameter information.

2. The software requirement document parsing method of claim 1, wherein, performing layout analysis on the demand document regarding text information and image information to obtain layout area information, specifically comprising: performing full-text scanning on the demand document through a preset image processing algorithm to obtain a full-text layout image; performing feature recognition on the full-text layout image based on an image feature recognition frame, and uniformly marking the recognized image features to obtain image marked features; The text features in the full-text layout image are recognized and marked by a preset OCR recognition technology to obtain a first marked region; According to the main features of the requirement document, the full-text layout image is divided into regions related to document content attributes, and a second marked region is marked; wherein the main features include document layout features, document title features, and document key marker features of the requirement document; the document content attributes include text attributes, title attributes, table attributes, and picture attributes of the requirement document; Based on the first marked region and the second marked region, the image marking features are classified and categorized according to the layout region, and the layout region information is determined.

3. The software requirement document parsing method of claim 2, wherein, Based on the first marked region and the second marked region, the image marking features are classified and categorized according to the layout region, and the layout region information is determined, specifically including: The first marked region and the image marked region are matched according to image region features, and the region that meets the feature matching is determined as a functional requirement region; The second marked region and the image marked region are matched according to image region features, and the region that meets the feature matching is determined as a functional requirement region and a page requirement region; According to the functional requirement region and the page requirement region, the full-text layout image is divided into layout regions to obtain the layout region information.

4. The software requirement document parsing method of claim 1, wherein, After positioning and classifying each element of the page requirement information to be recognized based on the deep learning of the convolutional neural network, the method further includes: Through an OCR recognition technology, the text embedding features in the page requirement information to be recognized are converted into text features to obtain editable text features; Based on the deep learning of the convolutional neural network, the text features are positioned, and the positioned text features are determined as the first key information; Wherein, the first key information includes the positioned text feature information and the positioning and classification information of each element.

5. The software requirement document parsing method of claim 1, wherein, Based on the first key information and the second key information, the requirement document is trained for layout recognition to obtain a comprehensive requirement model, specifically including: Based on the first key information, the requirement document is trained for page template generation to obtain a page requirement model; Based on the first key information, the requirement document is trained for functional template generation to obtain a functional requirement model; The page requirement model and the functional requirement model are associated with the same requirement document to determine the comprehensive requirement model.

6. The software requirement document parsing method of claim 5, wherein, The page template includes: page type elements, icon type elements, text feature elements, and table style elements; The functional requirement model includes: word order elements, word type elements, specific entity elements, syntax structure elements, and dependency relationship elements.

7. A software requirement document analysis apparatus characterized by comprising: The device includes: at least one processor; and a memory in communication with the at least one processor; wherein The memory stores instructions capable of being executed by the at least one processor, so that the at least one processor can execute the software requirement document analysis method according to any one of claims 1-6.

8. A non-transitory computer storage medium, comprising, The storage medium is a nonvolatile computer readable storage medium, and the nonvolatile computer readable storage medium stores at least one program, and each program includes instructions, which, when executed by a terminal, causes the terminal to execute the software requirement document analysis method according to any one of claims 1-6.

Images