Electronic business card generation method and system based on deep learning

Abstract

The invention discloses an electronic business card generation method and system based on deep learning, and belongs to the technical field of electronic business card generation, and the method comprises the steps: firstly obtaining original multi-source business card data, carrying out the data cleaning of the original multi-source business card data, and forming multi-source business card data; then forming multi-source business card text data according to the multi-source business card data, and performing text feature extraction on the multi-source business card text data based on deep learning to form electronic business card content generation data; inputting the electronic business card content generation data into an electronic business card content generation model to obtain electronic business card content information; and finally, obtaining user business card template information, and carrying out information fusion on the electronic business card content information and the user business card template information to obtain the electronic business card. By obtaining the original multi-source business card data, the generation efficiency of the electronic business card is improved; moreover, through deep learning, preprocessing, feature extraction and content information training are carried out on the multi-source business card data, and the quality of the electronic business card content is greatly improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of electronic business card generation, and specifically relates to an electronic business card generation method and system based on deep learning. Background Art

[0002] Currently, electronic business cards (eBusiness Cards) serve as a core vehicle for displaying a professional image for both businesses and individuals, serving as a tool for efficient communication and accurate information delivery. However, current eBusiness Card generation tools still suffer from several flaws, hindering their efficiency and content quality from meeting the demands of modern business scenarios. This has become a key bottleneck restricting business networking efficiency and commercial value creation.

[0003] Among them, in terms of the efficiency of electronic business card generation, existing electronic business card generation tools generally rely on users to manually enter basic fields such as company name, position, contact information, etc., and lack deep integration and docking with data in mainstream office systems and industry databases. Taking multinational companies as an example, employees need to repeatedly fill in Chinese and English names, multi-level department affiliations and global branch information in a multilingual environment. The production of a single business card can take more than 15 minutes, and when batches of team business cards are generated, the time taken will be greatly extended. Its generation efficiency is far from meeting the time limit requirements in business social networking.

[0004] In addition, most existing electronic business card generation tools only support manual input or importing Excel spreadsheets by users, lack real-time crawling and cross-validation of network data sources, and because the content input by users may not meet the requirements for generating electronic business cards, the generated electronic business cards may become a messy and meaningless accumulation of information, making the quality of the generated electronic business cards fail to meet the requirements of business social interaction.

[0005] Therefore, based on the above-mentioned shortcomings, how to provide an electronic business card generation method and system based on deep learning that can improve the generation efficiency of electronic business cards and optimize the content quality of electronic business cards has become an urgent problem to be solved. Summary of the Invention

[0006] The purpose of the present invention is to provide an electronic business card generation method based on deep learning, so as to solve the problems existing in the prior art of low electronic business card generation efficiency and poor content quality of the generated electronic business cards.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions:

[0008] In a first aspect, the present invention provides a method for generating an electronic business card based on deep learning, comprising:

[0009] Acquiring original multi-source business card data and performing data cleaning on the original multi-source business card data to form multi-source business card data, wherein the multi-source business card data includes image business card data and text business card data;

[0010] Extracting image text data from the image business card data, aligning the image text data with the text business card data to form multi-source business card text data, and performing text feature extraction on the multi-source business card text data based on deep learning to obtain multi-source business card text feature data to form electronic business card content generation data;

[0011] Obtaining a pre-trained electronic business card content generation model, inputting the electronic business card content generation data into the electronic business card content generation model, and obtaining electronic business card content information;

[0012] The user business card template information is obtained, and the electronic business card content information and the user business card template information are fused to obtain an electronic business card corresponding to the original multi-source business card data.

[0013] In one possible design, obtaining original multi-source business card data and performing data cleaning on the original multi-source business card data to form multi-source business card data includes:

[0014] Obtaining the personal profile information input by the user and using the personal profile information input by the user as the user source business card data;

[0015] Based on the user source business card data, an artificial intelligence algorithm is used to search for information in enterprise databases, academic databases, and social databases to obtain user career information, user skill description information, and user contact information corresponding to the personal profile information input by the user, and the user career information, the user skill description information, and the user contact information are used as network source business card data;

[0016] Integrating the user-source business card data and the network-source business card data to form original multi-source business card data;

[0017] Performing data format recognition on the original multi-source business card data, and dividing the original multi-source business card data into structured data and unstructured data;

[0018] Perform duplicate record determination on each data record in the structured data, retain the latest data record, and remove the remaining data records to form de-redundant structured data. At the same time, use a hash algorithm to determine the data similarity of each data record in the unstructured data, and perform similar text de-duplication on data records that exceed a preset similarity threshold to form de-redundant unstructured data.

[0019] Based on de-redundant structured data and de-redundant unstructured data, regular expressions are used to modify the data format and obtain multi-source business card data.

[0020] In one possible design, extracting image text data from the image business card data includes:

[0021] Using the Otsu algorithm, each image in the image business card data is converted into a black and white binary image;

[0022] Using Gaussian filtering method, denoising and sharpening are performed on each of the black and white binary images to obtain pre-processed image business card data;

[0023] Extracting initial image text data from the pre-processed image business card data based on optical character recognition;

[0024] Regular expressions are used to filter out invalid symbols in the initial image text data, and the initial image text data is segmented into paragraphs according to line breaks or punctuation marks to form standard image text information, so as to use the standard image text information as image text data.

[0025] In one possible design, image text data and text business card data are aligned to form multi-source business card text data, including:

[0026] Obtaining preset field mapping relationships and priority rules, and performing named entity recognition on the multi-source business card text data using the T5 PEGASUS model;

[0027] The text format of the identified entities is standardized based on the preset field mapping relationship, and the identified entities are classified into predefined entity categories according to the preset priority rules to complete data alignment and obtain multi-source business card text data;

[0028] Accordingly, based on deep learning, text feature extraction is performed on the multi-source business card text data to obtain multi-source business card text feature data to form electronic business card content generation data, which includes:

[0029] Input the multi-source business card text data after entity classification into the T5 PEGASUS model, extract the hidden state of the CLS tag of the last layer of the model, and perform pooling processing to obtain the main feature vector;

[0030] Obtaining key entity categories to extract key entities from each entity category in the multi-source business card text data;

[0031] The key entity and the main feature vector are feature-concatenated to obtain multi-source business card text feature data, and the multi-source business card text feature data is used as electronic business card content generation data.

[0032] In a possible design, a pre-trained electronic business card content generation model is obtained, including:

[0033] Acquire historical electronic business card content generation data, and divide the historical electronic business card content generation data into a training data set, a verification data set, and a test data set;

[0034] Taking the T5 PEGASUS model as the base model, the training data set is input into the T5 PEGASUS model, and the electronic business card content is used as the output result for pre-training to obtain a pre-electronic business card content generation model;

[0035] Using the validation data set, adjusting the model parameters of the pre-electronic business card content generation model, and using the test data set to evaluate the pre-electronic business card content generation model, and using the pre-electronic business card content generation model that passes the model evaluation as the electronic business card content generation model;

[0036] Accordingly, inputting the electronic business card content generation data into the electronic business card content generation model to obtain electronic business card content information includes:

[0037] Inputting the electronic business card content generation data into the electronic business card content generation model to obtain an output result of the electronic business card content generation model;

[0038] The output result of the electronic business card content generation model is used as the electronic business card content information.

[0039] In a possible design, after obtaining the electronic business card content information, the following is also included:

[0040] Verify the integrity and format of electronic business card content, and filter out redundant information.

[0041] A key entity category and a preset logical rule are obtained, and a semantic consistency check is performed on the fields corresponding to the key entity category using the preset logical rule.

[0042] In one possible design, obtaining user business card template information and fusing the electronic business card content information with the user business card template information to obtain an electronic business card corresponding to the original multi-source business card data input by the user includes:

[0043] Acquire user business card template information, and complete business card style design and layout design of the user business card template based on the user business card template information to generate the user business card template;

[0044] Generate a scannable QR code based on the electronic business card content information;

[0045] The electronic business card content information and the QR code are integrated and filled into the user business card template to obtain an electronic business card corresponding to the original multi-source business card data input by the user.

[0046] In a second aspect, the present invention provides an electronic business card generation system based on deep learning, which includes:

[0047] A data acquisition module, configured to acquire original multi-source business card data and perform data cleaning on the original multi-source business card data to form multi-source business card data, wherein the multi-source business card data includes image business card data and text business card data;

[0048] a feature extraction module, configured to extract image text data from the image business card data, align the image text data with the text business card data to form multi-source business card text data, and perform text feature extraction on the multi-source business card text data based on deep learning to obtain multi-source business card text feature data to form electronic business card content generation data;

[0049] A content information training module is used to obtain a pre-trained electronic business card content generation model, input the electronic business card content generation data into the electronic business card content generation model, and obtain electronic business card content information;

[0050] The business card generation module is used to obtain user business card template information and fuse the electronic business card content information with the user business card template information to obtain an electronic business card corresponding to the original multi-source business card data.

[0051] In a third aspect, the present invention provides an electronic device comprising a memory, a processor and a transceiver that are communicatively connected in sequence, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the deep learning-based electronic business card generation method as described in the first aspect or any possible design of the first aspect.

[0052] In a fourth aspect, the present invention provides a computer-readable storage medium having instructions stored thereon. When the instructions are run on a computer, the electronic business card generation method based on deep learning as described in the first aspect or any possible design of the first aspect is executed.

[0053] In a fifth aspect, the present invention provides a computer program product comprising instructions, which, when executed on a computer, causes the computer to execute the electronic business card generation method based on deep learning as described in the first aspect or any possible design of the first aspect.

[0054] Beneficial effect: The present invention provides an electronic business card generation method and system based on deep learning, which first obtains original multi-source business card data and performs data cleaning on the original multi-source business card data to form multi-source business card data, wherein the multi-source business card data includes image business card data and text business card data; then, image text data is extracted from the image business card data, and the image text data is aligned with the text business card data to form multi-source business card text data, and based on deep learning, text feature extraction is performed on the multi-source business card text data to obtain multi-source business card text feature data to form electronic business card content generation data; then, a pre-trained electronic business card content generation model is obtained, and the electronic business card content generation data is input into the electronic business card content generation model to obtain electronic business card content information; finally, user business card template information is obtained, and the electronic business card content information and the user business card template information are information-fused to obtain an electronic business card corresponding to the original multi-source business card data. By acquiring original multi-source business card data, the information limitations of electronic business card generation are improved, the user's time in inputting information is saved, and the efficiency of electronic business card generation is improved; moreover, multi-source business card data combines text data and image data, making the generated electronic business card information more accurate. In addition, through deep learning, these multi-source business card data are pre-processed, feature extracted and content information trained, and more critical electronic business card content information that is more in line with the requirements of electronic business card production is obtained as filler content, which greatly improves the quality of electronic business card content. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A schematic diagram of the steps of a method for generating an electronic business card based on deep learning provided by an embodiment of the present invention;

[0056] Figure 2 A schematic diagram of the functional structure of an electronic business card generation system based on deep learning provided by an embodiment of the present invention;

[0057] Figure 3 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0058] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the present invention will be briefly introduced below in conjunction with the drawings and the description of the embodiments or the prior art. Obviously, the following description of the structure of the drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.

[0059] It should be understood that although the terms "first," "second," etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element can be referred to as a second element, and similarly, a second element can be referred to as a first element without departing from the scope of the exemplary embodiments of the present invention.

[0060] It should be understood that the term "and / or" that may appear in this document is merely a description of the association relationship between associated objects, indicating that three relationships may exist. For example, A and / or B may indicate three situations: A exists alone, B exists alone, and A and B exist at the same time. The term " / and" that may appear in this document describes another type of association object relationship, indicating that two relationships may exist. For example, A / and B may indicate two situations: A exists alone, and A and B exist alone. In addition, the character " / " that may appear in this document generally indicates that the previous and subsequent associated objects are in an "or" relationship.

[0061] Example:

[0062] like Figure 1 As shown, the first aspect of this embodiment provides a method for generating an electronic business card based on deep learning, which includes:

[0063] S1. Obtaining original multi-source business card data and performing data cleaning on the original multi-source business card data to form multi-source business card data, wherein the multi-source business card data includes image business card data and text business card data;

[0064] In a possible implementation, in step S1, the process of obtaining original multi-source business card data and performing data cleaning on the original multi-source business card data to form multi-source business card data can be decomposed into, but not limited to, the following steps S11-S16, including:

[0065] S11 obtains the personal profile information entered by the user and uses the personal profile information entered by the user as the user's source business card data;

[0066] S12. Based on the user-source business card data, an artificial intelligence algorithm is used to search for information in enterprise databases, academic databases, and social databases to obtain the user's professional experience information, user skill description information, and user contact information corresponding to the personal profile information entered by the user, and the user's professional experience information, user skill description information, and user contact information are used as the network-source business card data;

[0067] S13 integrates the user source business card data and the network source business card data to form the original multi-source business card data;

[0068] S14. Identify the data format of the original multi-source business card data and divide the original multi-source business card data into structured data and unstructured data;

[0069] S15. Determine duplicate records for each data record in the structured data, retain the latest data record, and remove the remaining data records to form de-redundant structured data. At the same time, use a hash algorithm to determine the data similarity of each data record in the unstructured data, and remove duplicate text from data records that exceed a preset similarity threshold to form de-redundant unstructured data.

[0070] S16. Based on the de-redundant structured data and the de-redundant unstructured data, regular expressions are used to modify the data format to obtain multi-source business card data.

[0071] It should be noted that to obtain the personal profile information entered by the user, the user only needs to enter a personal profile containing information such as name and position, and no specific or complex information is required. This greatly saves the user's time in entering information and improves the efficiency of electronic business card generation from the source; at the same time, based on the personal profile information entered by the user, information search is performed in corporate databases, academic databases and social databases to obtain user career experience information, user skill description information and user contact information corresponding to the personal profile information entered by the user. This information includes text, images and structured tables, covering information such as career experience, skill description and contact information, and integrates the user source business card data and the network source business card data. This multi-source information acquisition method improves the information limitations of electronic business card generation; and the collected original data is cleaned, including but not limited to removing duplicate data, correcting format errors, unifying the coding format, etc., and the unstructured data is sentence and word segmented. Through these data processing, more accurate text data is obtained, making the generated electronic business card information more accurate.

[0072] S2. Extracting image text data from the image business card data, aligning the image text data with the text business card data to form multi-source business card text data, and performing text feature extraction on the multi-source business card text data based on deep learning to obtain multi-source business card text feature data to form electronic business card content generation data;

[0073] In a possible implementation, in step S2, the process of extracting image text data from the image business card data may be decomposed into, but not limited to, the following steps S21-S24, including:

[0074] S21 using the Otsu algorithm, the image business card data in each image is converted into a black and white binary image;

[0075] S22. Using Gaussian filtering, denoising is performed on each of the black and white binary images, and image sharpening is performed to obtain pre-processed image business card data;

[0076] S23 based on optical character recognition, extracting the initial image text data from the pre-processed image business card data;

[0077] S24. Using regular expressions, filter out invalid symbols in the initial image text data, and split the initial image text data into paragraphs according to line breaks or punctuation marks to form standard image text information, and use the standard image text information as image text data.

[0078] In a possible implementation, in step S2, the process of aligning the image text data with the text business card data to form multi-source business card text data may be decomposed into, but not limited to, the following steps S25-S26, including:

[0079] S25 obtains the preset field mapping relationship and priority rules, and uses the T5 PEGASUS model to perform named entity recognition on the multi-source business card text data;

[0080] S26. Standardize the text format of the identified entities based on the preset field mapping relationship, and classify the identified entities into predefined entity categories according to the preset priority rules, complete data alignment, and obtain multi-source business card text data;

[0081] Accordingly, based on deep learning, the process of performing text feature extraction on the multi-source business card text data to obtain multi-source business card text feature data to form electronic business card content generation data can be decomposed into, but not limited to, the following steps S27-S29, including:

[0082] S27. The multi-source business card text data after entity classification is input into the T5 PEGASUS model, the hidden state of the last layer of the model CLS tag is extracted, and pooling is performed to obtain the main feature vector;

[0083] S28 obtains key entity categories to extract key entities from each entity category in the multi-source business card text data;

[0084] S29. Perform feature concatenation on the key entity and the main feature vector to obtain multi-source business card text feature data, and use the multi-source business card text feature data as electronic business card content generation data.

[0085] It should be noted that each image in the image business card data is processed by first converting it into a grayscale image to reduce the computational complexity and then converting the image into a black and white binary image through the Otsu algorithm (Otsu threshold segmentation algorithm) to enhance the text contrast. Then, the Gaussian filter method is used to remove the noise in the binary image, and the image is sharpened to enhance the text edge for optical character recognition, so as to obtain the preprocessed image business card data. Named entity recognition can preferentially use the BERT-BiLSTM-CRF joint model, and for the field of electronic business cards, the predefined entity categories may include but are not limited to name, position, company, email, phone, address and skill labels.

[0086] S3 obtains a pre-trained electronic business card content generation model, inputs the electronic business card content generation data into the electronic business card content generation model, and obtains electronic business card content information;

[0087] In a possible implementation, in step S3, the process of obtaining a pre-trained electronic business card content generation model may be decomposed into, but not limited to, the following steps S31-S33, including:

[0088] S31 obtains historical electronic business card content generation data, and divides the historical electronic business card content generation data into a training data set, a validation data set, and a test data set;

[0089] S32. Using the T5 PEGASUS model as the base model, the training data set is input into the T5 PEGASUS model, and the electronic business card content is used as the output result for pre-training to obtain a pre-electronic business card content generation model;

[0090] S33 uses the validation data set to adjust the model parameters of the pre-electronic business card content generation model, and uses the test data set to evaluate the pre-electronic business card content generation model, and uses the pre-electronic business card content generation model evaluated by the model as the electronic business card content generation model;

[0091] Accordingly, the process of inputting the electronic business card content generation data into the electronic business card content generation model to obtain the electronic business card content information can be decomposed into, but not limited to, the following steps S34-S35, including:

[0092] S34. Input the electronic business card content generation data into the electronic business card content generation model to obtain the output result of the electronic business card content generation model;

[0093] S35. Use the output result of the electronic business card content generation model as the electronic business card content information.

[0094] In a possible implementation, after step S3, after obtaining the electronic business card content information, the following steps are further included:

[0095] Verify the integrity and format of electronic business card content, and filter out redundant information.

[0096] A key entity category and a preset logical rule are obtained, and a semantic consistency check is performed on the fields corresponding to the key entity category using the preset logical rule.

[0097] It should be noted that the T5 PEGASUS model is a generative pre-trained model for Chinese that integrates T5 and PEGASUS. It is highly capable of text generation tasks. Its strong small-sample learning capabilities and efficient model optimization make it particularly suitable for this embodiment and serve as the foundation of the e-card content generation model. The e-card content generation model is trained using historical e-card content generation data and dynamically adjusts each user's e-card content generation data. This precise e-card content generation model enables the generation of more critical e-card content that better meets e-card production requirements and serves as filler content in the e-card, significantly improving the quality of the e-card content.

[0098] S4. Obtain user business card template information, and fuse the electronic business card content information with the user business card template information to obtain an electronic business card corresponding to the original multi-source business card data.

[0099] In one possible implementation, in step S4, the process of obtaining user business card template information and fusing the electronic business card content information with the user business card template information to obtain an electronic business card corresponding to the original multi-source business card data input by the user may be decomposed into, but not limited to, the following steps S41-S43, including:

[0100] S41 obtains the user's business card template information, and completes the user's business card template style design and layout design based on the user's business card template information to generate a user's business card template;

[0101] S42. Generate a scannable QR code based on the electronic business card content information;

[0102] S43. Integrate the electronic business card content information and the QR code and fill in the user business card template to obtain an electronic business card corresponding to the original multi-source business card data input by the user.

[0103] It should be noted that the user business card template information includes information such as background image, corporate logo and text font. The business card style design of the user business card template can unify the color, transparent image and style of the user business card template to ensure visual consistency; the layout design of the user business card template is dynamically adaptively adjusted according to the specific electronic business card content information to obtain appropriate text placement position and size, image placement position and size and QR code placement position and size; the electronic business card content information and the QR code are integrated, and various text and image information are filled in the template using overlay layer technology, and finally an electronic business card corresponding to the original multi-source business card data input by the user is obtained. In addition, according to specific requirements, electronic business cards can be generated in PDF, HTML or image format to support user customized selection, so that users can use the generated electronic business cards in various business usage scenarios.

[0104] like Figure 2 As shown, the second aspect of this embodiment provides a hardware system for implementing the electronic business card generation method based on deep learning described in the first aspect of the embodiment, including:

[0105] A data acquisition module, configured to acquire original multi-source business card data and perform data cleaning on the original multi-source business card data to form multi-source business card data, wherein the multi-source business card data includes image business card data and text business card data;

[0106] a feature extraction module, configured to extract image text data from the image business card data, align the image text data with the text business card data to form multi-source business card text data, and perform text feature extraction on the multi-source business card text data based on deep learning to obtain multi-source business card text feature data to form electronic business card content generation data;

[0107] A content information training module is used to obtain a pre-trained electronic business card content generation model, input the electronic business card content generation data into the electronic business card content generation model, and obtain electronic business card content information;

[0108] The business card generation module is used to obtain user business card template information and fuse the electronic business card content information with the user business card template information to obtain an electronic business card corresponding to the original multi-source business card data.

[0109] The working process, working details and technical effects of the system provided in this embodiment can be found in the first aspect of the embodiment and will not be described in detail here.

[0110] like Figure 3As shown, the third aspect of this embodiment provides an electronic device, comprising: a memory, a processor and a transceiver that are communicatively connected in sequence, wherein the memory is used to store computer programs, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the electronic business card generation method based on deep learning as described in the first aspect of the embodiment.

[0111] For example, the memory may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, first-in first-out memory (FIFO), and / or first-in last-out memory (FILO); specifically, the processor may include one or more processing cores, such as a quad-core processor, an octa-core processor, etc. The processor may be implemented in at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). Furthermore, the processor may include a main processor and a coprocessor. The main processor is a processor for processing data in an awake state, also known as a CPU (Central Processing Unit); and the coprocessor is a low-power processor for processing data in a standby state.

[0112] In some embodiments, the processor may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. For example, the processor may be, but is not limited to, a microprocessor of the STM32F105 series, a reduced instruction set computer (RISC) microprocessor, an X86 architecture processor, or a processor with an integrated embedded neural network processing unit (NPU); the transceiver may be, but is not limited to, a wireless fidelity (WIFI) wireless transceiver, a Bluetooth wireless transceiver, a general packet radio service technology (GPRS) wireless transceiver, a ZigBee protocol (a low-power local area network protocol based on the IEEE802.15.4 standard, ZigBee) wireless transceiver, a 3G transceiver, a 4G transceiver, and / or a 5G transceiver. In addition, the device may also include, but is not limited to, a power module, a display screen, and other necessary components.

[0113] The working process, working details and technical effects of the electronic device provided in this embodiment can be found in the first aspect of the embodiment and will not be described in detail here.

[0114] The fourth aspect of this embodiment provides a storage medium that stores instructions for the electronic business card generation method based on deep learning as described in the first aspect of the embodiment, that is, the storage medium stores instructions, and when the instructions are run on a computer, the electronic business card generation method based on deep learning as described in the first aspect of the embodiment is executed.

[0115] The storage medium refers to a carrier for storing data, which may include but is not limited to a floppy disk, an optical disk, a hard disk, a flash memory, a USB flash drive and / or a memory stick, and the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0116] The working process, working details and technical effects of the storage medium provided in this embodiment can be found in the first aspect of the embodiment and will not be described in detail here.

[0117] The fifth aspect of this embodiment provides a computer program product containing instructions, which, when executed on a computer, enables the computer to execute the electronic business card generation method based on deep learning as described in the first aspect of the embodiment, wherein the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0118] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included in the scope of protection of the present invention.

Claims

1. A method for generating electronic business cards based on deep learning, characterized in that: include: Acquiring original multi-source business card data and performing data cleaning on the original multi-source business card data to form multi-source business card data, wherein the multi-source business card data includes image business card data and text business card data; Extracting image text data from the image business card data, aligning the image text data with the text business card data to form multi-source business card text data, and performing text feature extraction on the multi-source business card text data based on deep learning to obtain multi-source business card text feature data to form electronic business card content generation data; Obtaining a pre-trained electronic business card content generation model, inputting the electronic business card content generation data into the electronic business card content generation model, and obtaining electronic business card content information; The user business card template information is obtained, and the electronic business card content information and the user business card template information are fused to obtain an electronic business card corresponding to the original multi-source business card data.

2. The method for generating an electronic business card based on deep learning according to claim 1, characterized in that: Acquiring original multi-source business card data and performing data cleaning on the original multi-source business card data to form multi-source business card data includes: Obtaining the personal profile information input by the user and using the personal profile information input by the user as the user source business card data; Based on the user source business card data, an artificial intelligence algorithm is used to search for information in enterprise databases, academic databases, and social databases to obtain user career information, user skill description information, and user contact information corresponding to the personal profile information input by the user, and the user career information, the user skill description information, and the user contact information are used as network source business card data; Integrating the user-source business card data and the network-source business card data to form original multi-source business card data; Performing data format recognition on the original multi-source business card data, and dividing the original multi-source business card data into structured data and unstructured data; Perform duplicate record determination on each data record in the structured data, retain the latest data record, and remove the remaining data records to form de-redundant structured data. At the same time, use a hash algorithm to determine the data similarity of each data record in the unstructured data, and perform similar text de-duplication on data records that exceed a preset similarity threshold to form de-redundant unstructured data. Based on de-redundant structured data and de-redundant unstructured data, regular expressions are used to modify the data format and obtain multi-source business card data.

3. The method for generating an electronic business card based on deep learning according to claim 1, characterized in that: Extracting image text data from the image business card data includes: Using the Otsu algorithm, each image in the image business card data is converted into a black and white binary image; Using Gaussian filtering method, denoising and sharpening are performed on each of the black and white binary images to obtain pre-processed image business card data; Extracting initial image text data from the pre-processed image business card data based on optical character recognition; Regular expressions are used to filter out invalid symbols in the initial image text data, and the initial image text data is segmented into paragraphs according to line breaks or punctuation marks to form standard image text information, so as to use the standard image text information as image text data.

4. The method for generating an electronic business card based on deep learning according to claim 1, characterized in that: Aligning the image text data with the text business card data to form multi-source business card text data, including: Obtaining preset field mapping relationships and priority rules, and performing named entity recognition on the multi-source business card text data using the T5 PEGASUS model; The text format of the identified entities is standardized based on the preset field mapping relationship, and the identified entities are classified into predefined entity categories according to the preset priority rules to complete data alignment and obtain multi-source business card text data; Accordingly, based on deep learning, text feature extraction is performed on the multi-source business card text data to obtain multi-source business card text feature data to form electronic business card content generation data, which includes: Input the multi-source business card text data after entity classification into the T5 PEGASUS model, extract the hidden state of the CLS tag of the last layer of the model, and perform pooling processing to obtain the main feature vector; Obtaining key entity categories to extract key entities from each entity category in the multi-source business card text data; The key entity and the main feature vector are feature-concatenated to obtain multi-source business card text feature data, and the multi-source business card text feature data is used as electronic business card content generation data.

5. The method for generating an electronic business card based on deep learning according to claim 1, characterized in that: Get pre-trained vCard content generation models, including: Acquire historical electronic business card content generation data, and divide the historical electronic business card content generation data into a training data set, a verification data set, and a test data set; Taking the T5 PEGASUS model as the base model, the training data set is input into the T5 PEGASUS model, and the electronic business card content is used as the output result for pre-training to obtain a pre-electronic business card content generation model; Using the validation data set, adjusting the model parameters of the pre-electronic business card content generation model, and using the test data set to evaluate the pre-electronic business card content generation model, and using the pre-electronic business card content generation model that passes the model evaluation as the electronic business card content generation model; Accordingly, inputting the electronic business card content generation data into the electronic business card content generation model to obtain electronic business card content information includes: Inputting the electronic business card content generation data into the electronic business card content generation model to obtain an output result of the electronic business card content generation model; The output result of the electronic business card content generation model is used as the electronic business card content information.

6. The method for generating an electronic business card based on deep learning according to claim 1, characterized in that: After obtaining the electronic business card content information, it also includes: Verify the integrity and format of electronic business card content, and filter out redundant information. A key entity category and a preset logical rule are obtained, and a semantic consistency check is performed on the fields corresponding to the key entity category using the preset logical rule.

7. The method for generating an electronic business card based on deep learning according to claim 1, characterized in that: Acquiring user business card template information and fusing the electronic business card content information with the user business card template information to obtain an electronic business card corresponding to the original multi-source business card data input by the user, including: Acquire user business card template information, and complete business card style design and layout design of the user business card template based on the user business card template information to generate the user business card template; Generate a scannable QR code based on the electronic business card content information; The electronic business card content information and the QR code are integrated and filled into the user business card template to obtain an electronic business card corresponding to the original multi-source business card data input by the user.

8. An electronic business card generation system based on deep learning, characterized in that: include: A data acquisition module, configured to acquire original multi-source business card data and perform data cleaning on the original multi-source business card data to form multi-source business card data, wherein the multi-source business card data includes image business card data and text business card data; a feature extraction module, configured to extract image text data from the image business card data, align the image text data with the text business card data to form multi-source business card text data, and perform text feature extraction on the multi-source business card text data based on deep learning to obtain multi-source business card text feature data to form electronic business card content generation data; A content information training module is used to obtain a pre-trained electronic business card content generation model, input the electronic business card content generation data into the electronic business card content generation model, and obtain electronic business card content information; The business card generation module is used to obtain user business card template information and fuse the electronic business card content information with the user business card template information to obtain an electronic business card corresponding to the original multi-source business card data.

9. An electronic device, characterized in that: It includes a memory, a processor and a transceiver that are communicatively connected in sequence, wherein the memory is used to store computer programs, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the electronic business card generation method based on deep learning as described in any one of claims 1 to 7.

10. A computer program product comprising a computer program or instructions, characterized in that When the computer program or the instruction is executed by a computer, the electronic business card generation method based on deep learning according to any one of claims 1 to 7 is implemented.

Images