Self-service printer document automatic typesetting and printing system and method combined with visual recognition

Abstract

The present invention relates to the technical field of automatic printing systems, and in particular to a self-service printer document automatic typesetting and printing system and method combined with visual recognition. Through the cooperation of a user identity authentication module, a server and data management module, a visual recognition and typesetting module, a printing execution module and an interface interconnection module, the document content and format can be automatically identified in combination with visual recognition, which greatly reduces the time for users to manually set and adjust, so as to improve the printing efficiency of the document. Through the setting of the visual recognition and typesetting module, the content and format of the uploaded document can be automatically identified through visual recognition technology without the need for manual setting by the user. According to the recognition result, the system automatically performs typesetting optimization to ensure that the layout of the printed output document is reasonable and beautiful. This automated recognition and typesetting process greatly reduces the time spent by users on setting and adjustment, and improves the overall printing efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of automatic printing systems, and in particular relates to a self-service printer document automatic typesetting and printing system and method combined with visual recognition. Background Art

[0002] The self-service printer document automatic typesetting and printing system provides users with an efficient, convenient and secure modern printing solution. It greatly improves the traditional printing mode and meets the diverse market needs through multi-terminal access, automatic typesetting, cloud processing, data security and other technical means.

[0003] Currently, the self-service printer document automatic typesetting and printing system on the market is mainly composed of three parts: hardware, software and network. The hardware includes printers, paper lending systems and card readers. The software is used to manage the hardware resources of the printer and provide various services. The network is convenient for users to perform remote printing management. Through the organic combination of hardware, software and network, efficient, convenient and secure printing services are achieved.

[0004] However, the above-mentioned printing system usually relies on manual visual confirmation of document content and format during printing, which makes the user's manual setting and adjustment time-consuming, thereby affecting the printing efficiency of the document and increasing labor costs. Therefore, we need to propose a self-service printer document automatic typesetting and printing system and method combined with visual recognition to solve the above-mentioned problems, so that it can automatically identify the document content and format in combination with visual recognition, greatly reducing the user's manual setting and adjustment time to improve the printing efficiency of the document. Summary of the invention

[0005] In view of the above problems, the present invention provides a self-service printer document automatic typesetting and printing system combined with visual recognition, including a user identity authentication module capable of supporting multiple identity authentication methods, a server and data management module for managing printing tasks and processing data, a visual recognition and typesetting module capable of automatically identifying the content and format of uploaded documents, a printing execution module for performing printing operations, and an interface interconnection module for smooth data transmission;

[0006] The user identity authentication module sends the verification information to the server through a secure data transmission protocol. After receiving the verification information, the server performs verification processing and returns the result to the user module. If the verification is successful, the user can further use the printing service. When the user uploads documents to the server, the server data management module will initially process and store these files. Then, the server sends the document to the visual recognition and typesetting module for analysis and typesetting. After the visual recognition and typesetting module completes the recognition and typesetting of the document, it transfers the processed document to the printing execution module in a format suitable for printing. After receiving the document, the printing execution module controls the printer to perform specific printing operations.

[0007] Furthermore, the user identity authentication module includes a user terminal, an application system server, a real-name authentication server, a face recognition module and a result return unit. The user terminal is a terminal where the user inputs personal information and initiates an authentication request; the application system server receives the authentication request submitted by the user terminal and forwards it to the real-name authentication server;

[0008] The real-name authentication server processes the authentication request submitted by the application system server and guides the user to perform the biometric recognition operation of face recognition;

[0009] The face recognition module collects the user's face image through a camera device and compares it with the image stored in the database to verify the user's identity.

[0010] Furthermore, the server and data management module includes a self-service printing server capable of managing printing tasks and controlling the entire printing process, and a database manager for storing document templates. The self-service printing server is configured as a Windows Server server or a Linux server, and the database manager includes a relational database and a non-relational database.

[0011] Furthermore, the process of the visual recognition and typesetting module in automatically recognizing the content and format of the uploaded document is as follows:

[0012] A1. First, perform image enhancement, denoising and binarization operations on the uploaded document;

[0013] A2. Use convolutional neural network model to extract key features in the image;

[0014] A3. Use the Mask R-CNN algorithm to segment the document image into different regions and identify the type of each region. Use the Mask R-CNN algorithm to segment the document image into different regions and identify the type of each region.

[0015] A4. Use optical character recognition technology to extract characters from images and convert them into editable electronic document formats;

[0016] A5. Analyze the logical structure of the document to ensure that the converted electronic document maintains the original structure and format.

[0017] Furthermore, in step A1, when performing image enhancement, denoising and binarization operations, the specific process is as follows:

[0018] B1. Adjust the grayscale distribution of the image so that it is evenly distributed on all possible grayscale values;

[0019] B2. By applying the Gamma function correction to each pixel value of the image, the brightness of the image is adjusted to improve the visual effect. The value range of the Gamma function is between 0.2-0.8:

[0020] B3. Select images A and B taken under different lighting conditions in the same scene, where A represents the base color and B represents the mixed color. The lighting distribution is L=A / B. Calculate the enhanced image Image using the following formula:

[0021] Image=A / B;

[0022] B4. Use the filtering formula to reduce the random noise of image A and image B and obtain the image pixel value , the calculation formula is as follows:

[0023] , where x is the x-axis pixel position and y is the y-axis pixel position. is the standard deviation;

[0024] B5. Take the median (I(x,y)) in each pixel neighborhood to replace the image pixel value , where I(x, y) is the set of pixel values ​​in the pixel neighborhood;

[0025] B6. Reduce noise while keeping the edge clear and calculate the weight value W (i, j) of the neighborhood pixel. The calculation formula is as follows:

[0026] , where i and j are the locations of the neighboring pixels, and x and y are the locations of the current pixel. and are the standard deviations of spatial and color differences, respectively;

[0027] B7. Calculate and optimize W(i, j) to obtain the weight value W(x, y). The optimization formula is as follows:

[0028] ),in, is the learning rate, is the partial derivative of the loss function with respect to the weight, and m is the number of samples;

[0029] B8. Divide the image pixels into two categories: foreground and background. Select the image with obvious contrast between foreground and background for binarization calculation. The binarization calculation formula is as follows:

[0030] B(x,y) = , where T is the threshold.

[0031] Furthermore, in step A2, the steps of extracting key features from the image using the convolutional neural network model are as follows:

[0032] C1. Input the image in the form of three-dimensional data in the convolutional neural network model;

[0033] C2. Extract local features of the input image and calculate the output of the feature map , the calculation formula is as follows:

[0034] ,in, is the output of the kth feature, is the number of convolution kernels, is the weight value, is the input data, is the bias term, is the activation function;

[0035] C3, use multiple different convolution kernels in the same convolution layer to extract multiple features, so that each convolution kernel generates a feature map, and multiple feature maps can jointly represent more complex image features;

[0036] C4, the application of activation function introduces nonlinear characteristics, so that the network can learn more complex features; the expression of ReLU function is max(0, x);

[0037] C5, the feature map after convolution and activation function processing, is downsampled through the pooling layer;

[0038] C6. Through multiple layers of stacked convolutional layers and pooling layers, the convolutional neural network model can gradually extract higher-level and more abstract features;

[0039] C7, the fully connected layer maps the extracted features to the final output. The fully connected layer is implemented through matrix multiplication and bias terms. The formula is:

[0040] y=Wx+b,

[0041] Where y is the output vector, W is the weight matrix, x is the input vector, and b is the bias vector.

[0042] Furthermore, the print execution module includes an input interface for receiving print tasks from a computer or other device, a processor for converting input print data into executable instructions, a printing device for printing operations, a paper motion controller for controlling paper movement, and a network interface for network communication.

[0043] Furthermore, the print execution module includes an input interface for receiving print tasks from a computer or other device, a processor for converting input print data into executable instructions, a printing device for printing operations, a paper motion controller for controlling paper movement, and a network interface for network communication. The print task is received from the computer or other device through the input interface, and the data of the print task is transmitted to the printer interface in the form of a document or image file; the received data is first temporarily stored in the printer's cache, and the input print data is converted into executable instructions by the processor.

[0044] Furthermore, the interface interconnection module includes data transmission and communication protocols, physical layer connections, network topology, flow control and cache consistency, and the network topology is set to one or more of a NOC topology, a point-to-point topology, a bus topology and a network type topology.

[0045] Based on the self-service printer document automatic typesetting and printing system combined with visual recognition described above, the present invention also provides a self-service printer document automatic typesetting and printing method combined with visual recognition, including the following steps:

[0046] S1. The user confirms his / her identity through the identity verification module. After the system verifies the user, it authorizes the legitimate user to proceed to the next step;

[0047] S2. The user uploads the document to be printed through the operation interface, and the document is set to image format, text format or mixed format;

[0048] S3, the server data management module receives and stores the uploaded files, and performs preliminary file format check and optimization processing;

[0049] S4, the visual recognition and typesetting module automatically identifies the content and format of the uploaded document, uses machine learning and image processing technology to parse the content, and automatically optimizes the typesetting based on the recognition results;

[0050] S5. The server data management module adds the processed document to the print queue, performs task management and scheduling, and the user views the status of the print task through the operation interface;

[0051] S6. The printing execution module receives the typeset document from the server and controls the printer to perform specific printing operations.

[0052] The beneficial effects of the present invention are:

[0053] 1. The present invention can automatically identify the content and format of a document by combining visual recognition through the cooperation of a user identity authentication module, a server and data management module, a visual recognition and typesetting module, a printing execution module and an interface interconnection module, thereby greatly reducing the time for manual setting and adjustment by the user to improve the printing efficiency of the document.

[0054] 2. The present invention can automatically identify the content and format of uploaded documents through the setting of visual recognition and typesetting modules through visual recognition technology, without the need for manual settings by users. According to the recognition results, the system automatically performs typesetting optimization to ensure that the layout of the printed output documents is reasonable and beautiful. This automated recognition and typesetting process greatly reduces the time spent by users on settings and adjustments, and improves the overall printing efficiency.

[0055] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0057] Figure 1 A system block diagram according to an embodiment of the present invention is shown;

[0058] Figure 2 A flowchart showing the automatic recognition of the content and format of an uploaded document by the visual recognition and typesetting module according to an embodiment of the present invention is shown;

[0059] Figure 3 A flowchart according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] The embodiment of the present invention provides a self-service printer document automatic typesetting and printing system combined with visual recognition, such as Figure 1-2 As shown, it includes a user authentication module that can support multiple authentication methods, a server and data management module for managing printing tasks and processing data, a visual recognition and typesetting module that can automatically identify the content and format of uploaded documents, a printing execution module for performing printing operations, and an interface interconnection module for smooth data transmission;

[0062] The user identity authentication module sends the verification information to the server through a secure data transmission protocol (such as HTTPS or TLS). After receiving the verification information, the server performs verification processing and returns the result to the user module. If the verification is successful, the user can further use the printing service. When the user uploads the document to the server, the server data management module will initially process and store these files. Then, the server sends the document to the visual recognition and typesetting module for analysis and typesetting. After the visual recognition and typesetting module completes the recognition and typesetting of the document, it transmits the processed document to the printing execution module in a format suitable for printing. After receiving the document, the printing execution module controls the printer to perform specific printing operations. The printing operation includes controlling the printing mechanical parts and setting the printing parameters. The printing execution module communicates data with the printer through the interface interconnection module to ensure the correct execution of the printing task. The interface interconnection module not only transmits the printing data, but also transmits the printer's status information (such as paper out, ink cartridge empty, etc.) back to the server data management module.

[0063] Through the cooperation of user identity authentication module, server and data management module, visual recognition and typesetting module, print execution module and interface interconnection module, it can automatically identify document content and format by combining visual recognition, greatly reducing the time of user manual setting and adjustment, so as to improve document printing efficiency.

[0064] The user identity authentication module includes a user terminal, an application system server, a real-name authentication server, a face recognition module and a result return unit, wherein the user terminal is a terminal where the user enters personal information, such as name, ID number, etc., and initiates an authentication request.

[0065] The application system server receives the authentication request submitted by the user terminal and forwards it to the real-name authentication server. The application system server acts as a bridge to pass the user's request to the back-end processing system.

[0066] The real-name authentication server processes the authentication request submitted by the application system server and guides the user to perform biometric recognition operations for face recognition.

[0067] The face recognition module collects the user's face image through a camera device and compares it with the image stored in the database to verify the user's identity. The face recognition module uses biometric recognition technology to improve the accuracy and security of identity authentication. The face recognition module uses common knowledge technology in the field and will not be described in detail here.

[0068] The result return unit returns the authentication result to the user terminal, application system server and real-name authentication server step by step after completing all comparisons, and finally confirms whether the identity authentication is successful.

[0069] From the user terminal to the real-name authentication server, strict security measures are adopted at each step to ensure the authenticity of the user's identity, which not only protects the user's information security, but also provides a solid security guarantee for applications such as self-service printers. When selecting and deploying such identity authentication modules, their technical maturity, security, and compatibility with existing systems should be fully considered to ensure that their advantages are maximized.

[0070] The server and data management module includes a self-service printing server that can manage printing tasks and control the entire printing process, and a database manager for storing document templates. The self-service printing server is set as a Windows Server server or a Linux server. The database manager includes a relational database and a non-relational database. The self-service printing server and the database manager each perform their respective duties, and together constitute the core backend system of the self-service printer. The server is responsible for processing and forwarding printing tasks, while the database stores and manages related data, which helps to better deploy and maintain the self-service printing system and ensure its efficient and stable operation.

[0071] When managing print tasks and controlling the entire printing process, the Windows Server server adds a print service role through the server manager to configure the server as a print server. After installing the print role, each printer needs to be configured in detail in the printer and print queue management interface. The detailed configuration includes port settings, driver selection, paper size and print priority. The print queue is the core component for processing and tracking print jobs. It can view, re-queue, cancel or pause print jobs to ensure the orderly progress of print tasks. Set a suitable access control list (ACL) for the printer to control the access rights of different users and groups to the printer and protect the security of documents. Shared printers can be deployed to clients through group policies, so that printing resources can be centrally managed and printing policies can be uniformly configured.

[0072] When managing print tasks and controlling the entire printing process, the Linux server can quickly add network printers and perform various configurations on them, such as setting print options and managing print queues, in the CUPS web interface or through command line tools. CUPS supports multiple print task management functions, such as automatic retry, pause, and cancel print jobs to ensure efficient operation of print tasks. For enterprise users, integrating the Linux print server with AD can achieve user authentication and fine-grained print permission control. By writing scripts, more advanced print task automation can be achieved on Linux, such as periodically checking the print queue length, automatically pausing or restarting the printer, etc.

[0073] The relational database uses tables to store data and reads and writes data through structured query language. In the self-service printing system, the relational database is used to store user information, document templates and printing records, supports transaction processing and concurrent access control, and ensures data consistency and integrity.

[0074] The non-relational database adopts a flexible data structure to process a large number of small files and high concurrent access. In the self-service printing application, the non-relational database is used to quickly read and write temporary data.

[0075] The process of the visual recognition and typesetting module in automatically identifying the content and format of the uploaded document is as follows:

[0076] A1. First, perform image enhancement, denoising and binarization operations on the uploaded document to facilitate subsequent processing;

[0077] A2. Use the convolutional neural network (CNN) model to extract key features from the image, which are helpful for subsequent layout analysis and character recognition;

[0078] A3. Use the Mask R-CNN algorithm to segment the document image into different regions and identify the type of each region, such as text, image, table, etc.; Use the Mask R-CNN algorithm to segment the document image into different regions and identify the type of each region. This requires multiple steps, including image preprocessing, feature extraction, etc. The Mask R-CNN algorithm is common knowledge in the art and will not be described in detail here.

[0079] A4. Use optical character recognition (OCR) technology to extract characters from images and convert them into editable electronic document formats;

[0080] A5. Analyze the logical structure of the document, such as distinguishing the title, body, paragraphs, etc., to ensure that the converted electronic document maintains the original structure and format.

[0081] Among them, in step A1, when performing image enhancement, denoising and binarization operations, the specific process is as follows:

[0082] B1. By adjusting the grayscale distribution of the image so that it is evenly distributed across all possible grayscale values, the contrast is increased. This method is suitable for images that are overall dark or bright and can effectively improve the visual effect.

[0083] B2. By applying the Gamma function correction to each pixel value of the image, the brightness of the image is adjusted to improve the visual effect. The value range of the Gamma function is between 0.2-0.8:

[0084] B3. Select images A and B taken under different lighting conditions in the same scene, where A represents the base color and B represents the mixed color. The lighting distribution is L=A / B. Calculate the enhanced image Image using the following formula:

[0085] Image=A / B;

[0086] B4. Use the filtering formula to reduce the random noise of image A and image B and obtain the image pixel value , the calculation formula is as follows:

[0087] , where x is the x-axis pixel position and y is the y-axis pixel position. is the standard deviation;

[0088] B5. Take the median (I(x,y)) in each pixel neighborhood to replace the image pixel value , where I(x, y) is the set of pixel values ​​in the pixel neighborhood;

[0089] B6. Reduce noise while keeping the edge clear and calculate the weight value W (i, j) of the neighborhood pixel. The calculation formula is as follows:

[0090] , where i and j are the locations of the neighboring pixels, and x and y are the locations of the current pixel. and are the standard deviations of spatial and color differences, respectively;

[0091] B7. Calculate and optimize W(i, j) to obtain the weight value W(x, y). The optimization formula is as follows:

[0092] ),in, is the learning rate, is the partial derivative of the loss function with respect to the weight, and m is the number of samples;

[0093] B8. Divide the image pixels into two categories: foreground and background. Select the image with obvious contrast between foreground and background for binarization calculation. The binarization calculation formula is as follows:

[0094] B(x,y) = , where T is the threshold;

[0095] In step A2, the steps of extracting key features from the image using the convolutional neural network model are as follows:

[0096] C1. Input the image in the form of three-dimensional data in the convolutional neural network model;

[0097] C2. Extract local features of the input image and calculate the output of the feature map , the calculation formula is as follows:

[0098] ,in, is the output of the kth feature, is the number of convolution kernels, is the weight value, is the input data, is the bias term, is the activation function;

[0099] C3, use multiple different convolution kernels in the same convolution layer to extract multiple features, so that each convolution kernel generates a feature map, and multiple feature maps can jointly represent more complex image features;

[0100] C4. Applying activation functions introduces nonlinear characteristics, enabling the network to learn more complex features. The expression of the ReLU function is max(0, x).

[0101] C5. The feature map after convolution and activation function processing is downsampled through a pooling layer (such as maximum pooling), which helps to reduce the amount of computation and prevent overfitting. At the same time, the pooling operation highlights the most important features and improves the spatial invariance of the features.

[0102] C6. By stacking multiple convolutional layers and pooling layers, the convolutional neural network model can gradually extract higher-level and more abstract features, thereby capturing the complex structures and patterns in the image.

[0103] C7, the fully connected layer maps the extracted features to the final output, such as classification labels or regression values. The fully connected layer is implemented through matrix multiplication and bias terms, and the formula is

[0104] y=Wx+b,

[0105] Where y is the output vector, W is the weight matrix, x is the input vector, and b is the bias vector.

[0106] Through the setting of visual recognition and typesetting modules, the content and format of uploaded documents can be automatically identified through visual recognition technology without the need for manual settings by users. According to the recognition results, the system automatically optimizes the typesetting to ensure that the layout of the printed documents is reasonable and beautiful. This automated recognition and typesetting process greatly reduces the time users spend on settings and adjustments, and improves overall printing efficiency.

[0107] The print execution module includes an input interface for receiving a print task from a computer or other device, a processor for converting input print data into executable instructions, a printing device for printing operations, a paper motion controller for controlling paper movement, and a network interface for network communication. The print task is received from the computer or other device through the input interface, and the data of the print task is transmitted to the interface of the printer in the form of a document or image file; the received data is first temporarily stored in the cache of the printer, and the input print data is converted into executable instructions by the processor;

[0108] The printing device includes a dot matrix printer, an inkjet printer and a laser printer. The dot matrix printer relies on the print head to obtain current pulses from the driving circuit. The electromagnet driving coil generates a magnetic field to attract the armature, driving the print needle to hit the ribbon, and printing a dot matrix pattern on the printing paper. The inkjet printer controls the nozzle to squeeze the ink or heat to generate bubbles, and sprays the ink on the paper to form a pattern. The laser printer uses electronic imaging technology. The laser emitted by the laser transmitter scans on the toner drum to form a charge pattern. Then the charged part attracts the charged toner, which is pressed onto the paper through the transfer roller, and then the toner is fixed on the paper through the heating component.

[0109] The paper motion controller moves the paper at the set line spacing in the print head or print roller to ensure that the image is printed on the paper line by line in sequence. This component ensures that the paper passes through the printer smoothly and prints the image in the correct position.

[0110] The network interface allows multiple computers to share a printer via a local area network or WiFi. In this mode, print tasks are transmitted to the print server via the network, and then managed by the server queue and allocated to specific printing devices.

[0111] The interface interconnection module includes data transmission and communication protocols, physical layer connections, network topology structures, flow control and cache consistency, wherein the data transmission and communication protocols use a network interconnection architecture to connect multi-core CPUs, GPUs and FPGA processors and acceleration slots through high-bandwidth, low-latency communication channels; the physical layer connection is to install the bare chip DIE on the substrate, and to connect the bare chip pins to the substrate through an internal connection path; the network topology structure is set to one or more of a NOC topology structure, a point-to-point topology structure, a bus topology structure and a network topology structure; flow control and cache consistency include a flow control mechanism that can prevent congestion and loss during data transmission and a cache consistency protocol that can ensure that different processors see consistent shared data, thereby ensuring data integrity and reliability.

[0112] Based on the above-described self-service printer document automatic typesetting and printing system combined with visual recognition, the present invention also provides a self-service printer document automatic typesetting and printing method combined with visual recognition, such as Figure 3 As shown, the following steps are included:

[0113] S1. The user confirms his / her identity through the identity verification module, which supports multiple verification methods such as password, fingerprint recognition or facial recognition. After the system verifies the user, it authorizes the legitimate user to proceed to the next step;

[0114] S2. The user uploads the document to be printed through the operation interface, and the document is set to image format, text format or mixed format; the operation interface can be printed and uploaded through smart devices, such as mobile phones, computers, etc.

[0115] S3, the server data management module receives and stores the uploaded files, and performs preliminary file format check and optimization processing;

[0116] S4. The visual recognition and typesetting module automatically identifies the content and format of the uploaded document, uses machine learning and image processing technology to parse the content, and automatically optimizes the typesetting based on the recognition results to ensure that the layout of the printed document is reasonable and beautiful, and is adapted to different printing sizes and format requirements;

[0117] S5. The server data management module adds the processed document to the print queue, performs task management and scheduling, and ensures that the printing task is carried out in an orderly manner; the user views the status of the printing task through the operation interface, such as the queue status, estimated printing time and other information.

[0118] S6. The print execution module receives the typeset document from the server and controls the printer to perform specific printing operations. During the printing process, the print execution module also monitors the printing status in real time, such as paper supply, ink cartridge remaining, etc., and promptly provides feedback to the user or system administrator when any problem occurs.

[0119] Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent substitutions for some of the technical features therein; and these modifications or substitutions do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A self-service printer document automatic typesetting and printing system combined with visual recognition, characterized by: It includes a user authentication module that can support multiple authentication methods, a server and data management module for managing printing tasks and processing data, a visual recognition and typesetting module that can automatically identify the content and format of uploaded documents, a print execution module for performing printing operations, and an interface interconnection module for smooth data transmission; The user identity authentication module sends the verification information to the server through a secure data transmission protocol. After receiving the verification information, the server performs verification processing and returns the result to the user module. If the verification is successful, the user can further use the printing service. When the user uploads documents to the server, the server data management module will preliminarily process and store these files. Subsequently, the server sends the documents to the visual recognition and typesetting module for analysis and typesetting. After the visual recognition and typesetting module completes the recognition and typesetting of the document, it transmits the processed document to the printing execution module in a format suitable for printing. After receiving the document, the printing execution module controls the printer to perform specific printing operations; The process of the visual recognition and typesetting module in automatically identifying the content and format of the uploaded document is as follows: A1. First, perform image enhancement, denoising and binarization operations on the uploaded document; A2. Use convolutional neural network model to extract key features in the image; A3. Use the Mask R-CNN algorithm to segment the document image into different regions and identify the type of each region. Use the Mask R-CNN algorithm to segment the document image into different regions and identify the type of each region. A4. Use optical character recognition technology to extract characters from images and convert them into editable electronic document formats; A5. Analyze the logical structure of the document to ensure that the converted electronic document maintains the original structure and format; In step A1, when performing image enhancement, denoising and binarization operations, the specific process is as follows: B1. Adjust the grayscale distribution of the image so that it is evenly distributed on all possible grayscale values; B2. By applying the Gamma function correction to each pixel value of the image, the brightness of the image is adjusted to improve the visual effect. The value range of the Gamma function is between 0.2-0.8: B3. Select images A and B taken under different lighting conditions in the same scene, where A represents the base color and B represents the mixed color. The lighting distribution is L=A / B. Calculate the enhanced image Image using the following formula: Image=A / B; B4. Use the filtering formula to reduce the random noise of image A and image B and obtain the image pixel value , the calculation formula is as follows: , where x is the x-axis pixel position and y is the y-axis pixel position. is the standard deviation; B5. Take the median (I(x,y)) in each pixel neighborhood to replace the image pixel value , where I(x, y) is the set of pixel values ​​in the pixel neighborhood; B6. Reduce noise while keeping the edge clear and calculate the weight value W (i, j) of the neighborhood pixel. The calculation formula is as follows: , where i and j are the locations of the neighboring pixels, and x and y are the locations of the current pixel. and are the standard deviations of spatial and color differences, respectively; B7. Calculate and optimize W(i, j) to obtain the weight value W(x, y). The optimization formula is as follows: ),in, is the learning rate, is the partial derivative of the loss function with respect to the weight, and m is the number of samples; B8. Divide the image pixels into two categories: foreground and background. Select the image with obvious contrast between foreground and background for binarization calculation. The binarization calculation formula is as follows: B(x,y) = , where T is the threshold; In step A2, the steps of extracting key features from the image using the convolutional neural network model are as follows: C1. Input the image in the form of three-dimensional data in the convolutional neural network model; C2. Extract local features of the input image and calculate the output of the feature map , the calculation formula is as follows: ,in, is the output of the kth feature, is the number of convolution kernels, is the weight value, is the input data, is the bias term, is the activation function; C3, use multiple different convolution kernels in the same convolution layer to extract multiple features, so that each convolution kernel generates a feature map, and multiple feature maps can jointly represent more complex image features; C4, the application of activation function introduces nonlinear characteristics, so that the network can learn more complex features; the expression of ReLU function is max(0, x); C5, the feature map after convolution and activation function processing, is downsampled through the pooling layer; C6. Through multiple layers of stacked convolutional layers and pooling layers, the convolutional neural network model can gradually extract higher-level and more abstract features; C7, the fully connected layer maps the extracted features to the final output. The fully connected layer is implemented through matrix multiplication and bias terms. The formula is: y=Wx+b, Among them, y is the output vector, W is the weight matrix, x is the input vector, and b is the bias vector; The print execution module includes an input interface for receiving a print task from a computer, a processor for converting input print data into executable instructions, a printing device for a print operation, a paper motion controller for controlling paper movement, and a network interface for network communication; The printing execution module includes an input interface for receiving printing tasks from a computer, a processor for converting input printing data into executable instructions, a printing device for printing operations, a paper motion controller for controlling paper movement, and a network interface for network communication. The printing task is received from a computer or other device through the input interface, and the data of the printing task is transmitted to the interface of the printer in the form of a document or image file; the received data is first temporarily stored in the cache of the printer, and the input printing data is converted into executable instructions by the processor.

2. The self-service printer document automatic typesetting and printing system combined with visual recognition according to claim 1 is characterized by: The user identity authentication module includes a user terminal, an application system server, a real-name authentication server, a face recognition module and a result return unit. The user terminal is where the user enters personal information and initiates an authentication request; The application system server receives the authentication request submitted by the user terminal and forwards it to the real-name authentication server; The real-name authentication server processes the authentication request submitted by the application system server and guides the user to perform the biometric recognition operation of face recognition; The face recognition module collects the user's face image through a camera device and compares it with the image stored in the database to verify the user's identity.

3. The self-service printer document automatic typesetting and printing system combined with visual recognition according to claim 2 is characterized by: The server and data management module includes a self-service printing server capable of managing printing tasks and controlling the entire printing process, and a database manager for storing document templates. The self-service printing server is configured as a Windows Server server or a Linux server, and the database manager includes a relational database and a non-relational database.

4. The self-service printer document automatic typesetting and printing system combined with visual recognition according to claim 3 is characterized by: The interface interconnection module includes data transmission and communication protocols, physical layer connections, network topology, flow control and cache consistency, and the network topology is set to one or more of a NOC topology, a point-to-point topology, a bus topology and a network topology.

5. A method for automatic typesetting and printing of documents by a self-service printer combined with visual recognition, based on a system for automatic typesetting and printing of documents by a self-service printer combined with visual recognition as claimed in any one of claims 1 to 4, characterized in that: The steps include: S1. The user confirms his / her identity through the identity verification module. After the system verifies the user, it authorizes the legitimate user to proceed to the next step; S2. The user uploads the document to be printed through the operation interface, and the document is set to image format, text format or mixed format; S3, the server data management module receives and stores the uploaded files, and performs preliminary file format check and optimization processing; S4, the visual recognition and typesetting module automatically identifies the content and format of the uploaded document, uses machine learning and image processing technology to parse the content, and automatically optimizes the typesetting based on the recognition results; S5. The server data management module adds the processed document to the print queue, performs task management and scheduling, and the user views the status of the print task through the operation interface; S6. The printing execution module receives the typeset document from the server and controls the printer to perform specific printing operations.

Images