Web table data entry method and system based on character recognition

Abstract

The invention provides a data entry method and system of a Web table based on character recognition, and relates to the technical field of image processing. The method comprises the following steps: acquiring to-be-input image data; constructing a Web table data set according to the image data; according to the Web table data set, configuring filling fields and table styles of the Web table; selecting an anchoring field and an identification area according to a filling field and a form style of the Web form; selecting a corresponding Web table text recognition plug-in according to a filling field and a table style of the Web table; extracting text content from the recognition area of the image data through a Web table text recognition plug-in; and automatically filling the text content into corresponding fields of the Web table according to the anchoring fields to generate a filling form. According to the method, the time for manually inputting the text can be shortened, and the text inputting efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly to a data entry method and system for Web tables based on character recognition. Background Art

[0002] Character recognition refers to extracting editable text content from pictures, scanned documents, or handwritten text through image processing technology. A Web table refers to a table presented on a website for displaying and inputting data. Data entry refers to the process of inputting data into a system or table, and data entry for Web tables based on character recognition refers to designing a new data entry method through character recognition technology, which can extract text information from images (such as scanned documents, screenshots, photos, etc.) and then automatically fill the extracted text into the Web table.

[0003] The data entry method and system for Web tables based on character recognition can automatically extract text information from images and fill it into Web tables, which can greatly improve the efficiency, accuracy, and automation of data entry, significantly improve the performance of Web tables in data entry, provide a high-efficiency, accurate, and easy-to-use data processing platform for users, and also bring higher operational efficiency and stronger market competitiveness to enterprises and organizations.

[0004] However, in existing Web table solutions, when users need to input non-text data, especially information extracted from images, they often rely on manual entry or additional development of specific functions to meet the requirements. This approach is not only inefficient but also prone to introducing human errors, increasing the cost and complexity of data processing, further leading to a poor user experience, and restricting the application scope of Web tables, especially showing obvious deficiencies when dealing with image data and other complex data types. There is an urgent need for a more intelligent and efficient solution to improve this situation. Summary of the Invention

[0005] In order to solve the technical problems in existing Web table solutions, which often rely on manual entry or additional development of specific functions to meet the requirements. This approach is not only inefficient but also prone to introducing human errors, increasing the cost and complexity of data processing, further leading to a poor user experience, and restricting the application scope of Web tables, especially showing obvious deficiencies when dealing with image data and other complex data types, the present invention provides a data entry method and system for Web tables based on character recognition.

[0006] The technical solutions provided by the embodiments of the present invention are as follows:

[0007] In the first aspect

[0008] A data entry method for a Web table based on text recognition provided by an embodiment of the present invention includes:

[0009] S1: Obtain the image data to be entered;

[0010] S2: Construct a Web table data set according to the image data;

[0011] S3: Configure the filling fields and table styles of the Web table according to the Web table data set;

[0012] S4: Select the anchoring fields and recognition areas according to the filling fields and table styles of the Web table;

[0013] S5: Select the corresponding Web table text recognition plug-in according to the filling fields and table styles of the Web table;

[0014] S6: Extract the text content from the recognition area through the Web table text recognition plug-in;

[0015] S7: Automatically fill the text content into the corresponding fields of the Web table according to the anchoring fields to generate a filled form.

[0016] In a second aspect

[0017] A data entry system for a Web table based on text recognition provided by an embodiment of the present invention includes:

[0018] A processor;

[0019] A memory, on which computer-readable instructions are stored. When the computer-readable instructions are executed by the processor, the data entry method for a Web table based on text recognition as described in the first aspect is implemented.

[0020] In a third aspect

[0021] A computer-readable storage medium provided by an embodiment of the present invention, on which a computer program is stored. When the program is executed by the processor, the data entry method for a Web table based on text recognition as described in the first aspect is implemented.

[0022] The beneficial effects brought by the technical solution provided by the embodiment of the present invention at least include:

[0023] In an embodiment of the present invention, by obtaining image data to be entered, constructing a Web table data set according to the image data, and configuring the filling fields and table styles of the Web table according to the Web table data set, it is ensured that the system can accurately fill in the data. According to the filling fields and table styles of the Web table, anchor fields and recognition regions are selected. Further, according to the filling fields and table styles of the Web table, a corresponding Web table text recognition plugin is selected. Through the Web table text recognition plugin, text content is extracted from the recognition region of the image data, and the text content is automatically filled into the corresponding fields of the Web table according to the anchor fields, generating a filled form. By automating the steps, the data entry efficiency is greatly improved, human errors are reduced, costs are lowered, and image data in different formats can be processed, thereby improving the processing ability and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0025] Figure 1 It is a flowchart showing a method for data entry of a Web table based on text recognition provided by an embodiment of the present invention;

[0026] Figure 2 It is a structural diagram showing a data entry system of a Web table based on text recognition provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0027] The following will describe the technical solutions in the present invention with reference to the drawings.

[0028] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to represent examples, illustrations, or explanations. Any embodiment or design solution described as an "example" in the present invention should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, the use of the word "example" is intended to present concepts in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one of the two can be selected.

[0029] To make the technical problems to be solved, technical solutions, and advantages of the present invention clearer, the following will be described in detail with reference to the drawings and specific embodiments.

[0030] Referring to the attached Figure 1 specification, it shows a flowchart of a method for data entry of a Web table based on text recognition provided by an embodiment of the present invention.

[0031] An embodiment of the present invention provides a method for data entry of a Web table based on character recognition. This method can be implemented by a data entry device for a Web table based on character recognition, and the data entry device for a Web table based on character recognition can be a terminal or a server. The processing flow of the data entry method for a Web table based on character recognition may include the following steps:

[0032] S1: Obtain image data to be entered.

[0033] Herein, "to be entered" means that the text or information contained in these image data needs to be extracted and input into a certain system or table. Image data refers to a digital image file containing text or graphic information obtained by means of scanning, photographing, or screenshotting, etc.

[0034] It should be noted that by obtaining image data, various types of data sources can be flexibly processed, and it is applicable to image files of different formats and sources. Whether it is a scanned paper document or an image taken by a mobile phone, camera, or other devices, it can be used as an input source, ensuring that the system can receive various different image data formats, thus achieving wide applicability.

[0035] In a possible implementation manner, the image data specifically includes: local image files, screenshot files, and scanned files.

[0036] It should be noted that automated acquisition of image data can save time, reduce manual entry errors, lay a foundation for subsequent character recognition and form filling, and improve the efficiency and accuracy of the entire data entry process.

[0037] S2: Construct a Web table data set according to the image data.

[0038] Herein, the Web table data set refers to a structural data set used to configure and display a Web table. It includes field information of the table, such as field names (e.g., name, address), field types (text, date, etc.), and field lengths, etc., to ensure that the table can correctly present and enter data.

[0039] It should be noted that by constructing a Web table data set, the field information of the Web table is reasonably configured and structured. The advantage is that it lays a foundation for subsequent automated data entry. By clearly defining field names, field types, and field lengths, the system can understand the characteristics and limitations of each table field, thereby avoiding input errors or format mismatch problems.

[0040] In a possible implementation manner, the Web table data set specifically includes: business field names, business field types, and business field lengths.

[0041] Among them, the business field name refers to the specific field name in the Web form, representing the data type or information content to be entered in each field, such as "Name", "Phone Number", etc.

[0042] Among them, the business field type refers to the data type expected by the form field, such as text, number, date, etc. This helps to determine the input format and validation rules of the field.

[0043] Among them, the business field length represents the maximum length of data that can be entered in the field. For example, the maximum number of characters for an "Address" field can be set to 255.

[0044] It should be noted that the Web form dataset provides consistency and standardization for the form, ensuring that form filling meets the requirements, and improving the accuracy and consistency of data entry. In addition, the clear data structure makes the subsequent text recognition and filling process more efficient, reducing error handling and adjustment in the subsequent stages of the system.

[0045] S3: Configure the filling fields and form styles of the Web form according to the Web form dataset.

[0046] Among them, the filling fields of the Web form refer to the specific areas or items used to enter data in the Web form, such as "Name", "Address", etc. Each field represents a place where user input can be accepted. The form style refers to the visual design and layout of the Web form, including field arrangement, color, font size, etc., aiming to improve the user experience and make the form clear, easy to understand and fill out.

[0047] It should be noted that by configuring the filling fields and form styles of the Web form, it is ensured that the form can be effectively presented to the user and provide clear operation guidelines. It not only guarantees the functionality of the form, but also improves the friendliness of the user interface. By reasonably configuring the fields and styles, users can easily identify which are required fields and which are optional fields, thus reducing operation errors and improving filling efficiency. At the same time, the optimization of the form style can ensure that the form looks neat and beautiful, enhancing the comfort of users when filling out and reducing the confusion of users during the filling process.

[0048] S4: Select the anchor fields and recognition areas according to the filling fields and form styles of the Web form.

[0049] Among them, the anchor field refers to the field in the Web form that corresponds to a specific area in the image. During the data entry process, the anchor field helps the system determine which form field the text extracted from the image should be filled into, ensuring that the data is accurately mapped to the corresponding position. The recognition area is the part of the image that is selected for text recognition, usually the area containing text content.

[0050] It should be noted that by selecting the anchoring fields and recognition regions, it is ensured that the text information extracted from the image can be correctly mapped to the corresponding fields of the Web table. The advantage of this step is that it can improve the accuracy and automation of data filling. By setting the anchoring fields, the system can identify and match the relationship between specific regions in the image and the Web table fields, reducing the error of manual input. The precise selection of the recognition region ensures the accuracy of text extraction, helps to avoid the extraction of irrelevant data, and ensures the correctness and integrity of the automatic filling of the table in subsequent steps.

[0051] S5: Select the corresponding Web table text recognition plug-in according to the filling fields and table styles of the Web table.

[0052] Among them, the Web table text recognition plug-in refers to a tool or software module used to extract text from an image. This plug-in can recognize the characters and text in the image and convert them into an editable text format for filling into the corresponding fields in the Web table.

[0053] It should be noted that by selecting a suitable Web table text recognition plug-in, it is ensured that the system can accurately and quickly extract text information from the image. Its advantage is that it enables the data in the image to be quickly converted into machine-readable text, avoiding the cumbersome process and errors of manual input. The correctly selected plug-in will use appropriate recognition algorithms for processing according to the image quality, text type, and format, thereby improving the accuracy and efficiency of recognition.

[0054] Specifically, the Web table text recognition plug-in usually implements the recognition of characters and text from an image based on optical character recognition (OCR) technology and converts them into an editable text format. The image will first be passed into the OCR plug-in for preprocessing. Then, the plug-in automatically detects the regions containing text in the image through algorithms (such as edge detection or convolutional neural networks). After the recognition regions are determined, the plug-in further splits these regions into individual characters. Character segmentation is completed by analyzing the spacing and shape between each character. For example, it will split the 'H', 'e', 'l', 'l', 'o' in 'Hello' apart. The plug-in uses a trained machine learning model to recognize each segmented character, and the recognized characters will be recombined into complete words and sentences. The plug-in will splice the recognized text according to the position and spatial relationship of the characters. Finally, the recognized text will be converted into a standard editable text format, such as a TXT file, JSON format, or directly filled into the input box of the Web table. At this time, the text can be used for further processing or editing, and the user does not need to input manually.

[0055] S6: Extract the text content from the recognition region through the Web table text recognition plug-in.

[0056] Among them, the text content refers to the text information recognized and extracted from the image. This information will be used to fill in the corresponding fields in the Web form.

[0057] It should be noted that the text content is extracted from the recognition area of the image data through the Web form text recognition plug-in. It automates the conversion process from image to text, saving a large amount of manual input time. The text recognition plug-in can accurately recognize the text in the specified area of the image, ensuring the accuracy of the extraction result.

[0058] In a possible implementation manner, extracting the text content from the recognition area of the image data in S6 specifically includes:

[0059] S601: Preprocess the recognition area through an edge detection algorithm to generate an edge image.

[0060] Among them, edge detection is an image processing technology used to identify edges or contours in an image. By detecting the places where the pixel values in the image change significantly through an algorithm, it is often used to extract important structures in the image, such as text, object boundaries, etc.

[0061] S602: Extract the column feature sequence of the edge image through the HOG descriptor.

[0062] Among them, the HOG descriptor is a technology used to describe the local shape features of an image and is often used in object detection. It extracts the edge and texture information of the image by calculating the histogram of gradient directions in the local area. This technology is widely used in fields such as character recognition and face recognition. The column feature sequence refers to the representation form of each column (or each small area) in the image in the feature space.

[0063] S603: Process the column feature sequence through a bidirectional LSTM and output a feature vector.

[0064] Among them, the bidirectional LSTM is an improved recurrent neural network (RNN) that can process sequence data in two directions (forward and backward). The feature vector is a numerical representation extracted from the data and contains the core information of the data. In this context, the feature vector is extracted from the column feature sequence through the LSTM model and can effectively represent the features of each character or text in the image.

[0065] S604: Extract the text content according to the feature vector, in combination with CTC and the vocabulary.

[0066] Among them, CTC is a loss function for sequence-to-sequence learning, especially suitable for dealing with input-output sequences of variable lengths. In character recognition, CTC can help the model generate the corresponding character sequence according to the extracted feature vectors, solving the alignment problem. The vocabulary is a set containing all possible output characters or words.

[0067] It should be noted that through this step, the key information (such as name, address, etc.) contained in the image can be quickly extracted and converted into text, reducing the risk of human errors and misunderstandings. In addition, the precise selection of the recognition area ensures that the system only processes the text related to the input, thereby improving the efficiency of the entire process and the accuracy of data processing.

[0068] In a possible implementation manner, S601 specifically includes:

[0069] S6011: Convert the color image in the recognition area into a grayscale image:

[0070] I gray I(x,y) = 0.2989×I R (x,y)+0.5870×I G (x,y)+0.1140×I B (x,y)

[0071] Among them, I gray (x,y) represents the grayscale value of the grayscale image at the position (x,y), and I R (x,y) represents the red channel value of the color image at the position (x,y), and I G (x,y) represents the green channel value of the color image at the position (x,y), and I B (x,y) represents the blue channel value of the color image at the position (x,y).

[0072] S6012: Smooth the grayscale image through a Gaussian filter:

[0073]

[0074] Among them, G(x,y) represents the value of the Gaussian function at the position (x,y), σ represents the standard deviation of the Gaussian function, π represents the pi, and exp represents the exponential function.

[0075] S6013: Based on the smoothed grayscale image, generate a smoothed image through a convolution operation:

[0076]

[0077] Among them, I smooth(x, y) represents the value of the smoothed image at position (x, y), i represents the offset in the horizontal direction, j represents the offset in the vertical direction, k represents the range of the smoothing operation, and I gray (x + i, y + j) represents the pixel value of the grayscale image at position (x + i, y + j).

[0078] S6014: Process the smoothed image through the Canny edge detection algorithm to generate an edge image.

[0079] It should be noted that the advantage of the Canny edge detection algorithm is that it can effectively and accurately extract the edges in the image, has strong anti-noise ability, and can provide continuous and accurate edge positions. These characteristics make it particularly suitable for preprocessing and feature extraction in image processing, and can provide high-quality image information for subsequent tasks such as text recognition and shape recognition.

[0080] In a possible implementation, S602 specifically includes:

[0081] S6021: Combine Edge Boxes and the ACF detector to identify candidate bounding boxes for text regions in the edge image.

[0082] Among them, Edge Boxes is an object detection method based on edge information, which generates candidate bounding boxes by analyzing the edges in the image. The ACF (Aggregate Channel Features) detector is a method for feature extraction and object detection based on image channels (such as gradient, color channels, etc.). A candidate bounding box for a text region refers to a region box that may contain text generated by an algorithm (such as Edge Boxes and the ACF detector).

[0083] S6022: Classify the candidate bounding boxes for text regions through a random forest to filter out the proposed bounding boxes in the candidate bounding boxes for text regions.

[0084] Among them, a random forest is an ensemble learning method that performs classification and regression tasks by constructing multiple decision trees and combining their outputs. Here, it is used to classify the candidate bounding boxes for text regions to filter out the bounding boxes that may contain text. A proposed bounding box is a candidate bounding box that may contain text after filtering.

[0085] S6023: Perform regression optimization on the proposed bounding boxes.

[0086] S6024: According to the proposed bounding boxes after regression optimization, extract the column feature sequence through the HOG descriptor:

[0087] C = Concat(HOG avg (i, j), for each column)

[0088]

[0089] Among them, C represents the column feature sequence, Concat represents the concatenation operation, and HOG avg (i, j) represents the average HOG feature value at position (i, j), T represents the size of the pooling window, p represents the index traversed during the pooling process, and HOG(p, j) represents the HOG feature at position (p, j).

[0090] It should be noted that first, Edge Boxes and the ACF detector are used to identify candidate boxes that may contain text in the image. Then, the candidate boxes are classified by a random forest to filter out the boxes containing text. Next, regression optimization is performed on these proposed boxes to make them more accurate. Finally, the column feature sequence in the optimized boxes is extracted using the HOG descriptor to provide the necessary feature information for subsequent character recognition. This process ensures the precise localization of the text area and effective feature extraction, providing a reliable basis for text recognition.

[0091] In a possible implementation manner, S603 is specifically:

[0092] According to the following formula, the column feature sequence is processed through a bidirectional LSTM to output a feature vector:

[0093] S t = S t-1 · f(W forget C t ) + f(W in C t )· f(W c C t )

[0094] Y t = f(W out C t )· g(S t )

[0095] Among them, S t represents the memory state at time t, S t-1 represents the memory state at time t - 1, f() represents the activation function, W forget represents the weight matrix of the forgetting gate, C t represents the column feature sequence at time t, W in represents the weight matrix of the input gate, W c represents the weight matrix of the candidate memory cell, Y t represents the feature vector at time t, W out represents the weight matrix of the output gate, and g() represents the tanh function.

[0096] Specifically, the bidirectional LSTM is a powerful sequence processing model, especially suitable for processing data with time-dependent relationships. It can learn information in the sequence from both the forward and backward directions simultaneously. Therefore, compared with the traditional LSTM, it can capture the dependencies between contexts more comprehensively.

[0097] It should be noted that the bidirectional LSTM can better understand and capture context information by processing sequence data in two directions (from past to future and from future to past). This means that the model can not only make judgments based on previous information but also make more accurate predictions based on subsequent information. For the recognition of text data, especially the sequence and arrangement of characters, the bidirectional LSTM provides more abundant information.

[0098] In one possible implementation, S604 specifically includes:

[0099] S6041: According to the feature vector, through CTC, calculate the probabilities of all output paths:

[0100]

[0101] Among them, p(W|C) represents the conditional probability of the output sequence W when the input sequence C is given, V(W) represents the set of all paths that can be mapped to the target word W, and p(π|C) represents the probability of the path π when the input sequence C is given.

[0102] S6042: Determine the loss function of CTC according to the probabilities:

[0103]

[0104] Among them, O represents the total value of the CTC loss function, S represents the training set, and log represents the logarithmic function.

[0105] S6043: With the goal of minimizing the loss function, combined with the vocabulary, select the best output path:

[0106] w * =argmax w∈L p(w|Y)

[0107] Among them, w * represents the optimal prediction result, argmax represents maximizing, L represents the vocabulary set, and p(w|Y) represents the conditional probability of the sequence w appearing when the probability matrix Y of the given output is given.

[0108] S6044: Extract the text content according to the best output path.

[0109] Among them, CTC (Connectionist Temporal Classification) is a loss function used to handle sequence-to-sequence learning tasks, especially suitable for cases where the lengths of the input and output sequences are not aligned.

[0110] It should be noted that the advantage of CTC is that it can effectively solve the problem of misalignment between the lengths of the input and output sequences. By calculating the probabilities of all possible paths and optimizing the loss function, it gradually improves the prediction accuracy. Combined with the vocabulary and decoding optimization, it can provide more accurate output and is suitable for automated text recognition tasks.

[0111] The beneficial effects brought by the technical solution provided by the embodiments of the present invention at least include:

[0112] In the embodiments of the present invention, by obtaining the image data to be entered, constructing a Web form dataset according to the image data, and configuring the filling fields and form styles of the Web form according to the Web form dataset, it is ensured that the system can accurately fill in the data. According to the filling fields and form styles of the Web form, the anchoring fields and recognition areas are selected. Further, according to the filling fields and form styles of the Web form, the corresponding Web form text recognition plug-in is selected. Through the Web form text recognition plug-in, the text content is extracted from the recognition area of the image data, and the text content is automatically filled into the corresponding fields of the Web form according to the anchoring fields, generating a filled form. The data entry efficiency is greatly improved through automated steps, human errors are reduced, costs are lowered, and image data in different formats can be processed, thereby improving the processing ability and accuracy.

[0113] Refer to the attached Figure 2 description, which shows a schematic structural diagram of a data entry system for a Web form based on text recognition provided by the present invention.

[0114] The present invention also provides a data entry system 20 for a Web form based on text recognition, which is applied to the above-mentioned data entry method for a Web form based on text recognition, including:

[0115] A processor 201.

[0116] A memory 202, on which computer-readable instructions are stored. When the computer-readable instructions are executed by the processor 201, the data entry method for a Web form based on text recognition as in the method embodiment is implemented.

[0117] The data entry system 20 for a Web form based on text recognition provided by the present invention can execute the above-mentioned data entry method for a Web form based on text recognition and achieve the same or similar technical effects. To avoid repetition, the present invention will not elaborate further.

[0118] The beneficial effects brought by the technical solution provided in the embodiment of the present invention at least include:

[0119] In the embodiment of the present invention, by obtaining the image data to be entered, constructing a Web table data set according to the image data, and configuring the filling fields and table styles of the Web table according to the Web table data set, it is ensured that the system can accurately fill in the data. According to the filling fields and table styles of the Web table, anchor fields and recognition areas are selected. Further, according to the filling fields and table styles of the Web table, the corresponding Web table text recognition plug-in is selected. Through the Web table text recognition plug-in, the text content is extracted from the recognition area of the image data, and the text content is automatically filled into the corresponding fields of the Web table according to the anchor fields, generating a filled form. The data entry efficiency is greatly improved through automated steps, human errors are reduced, costs are lowered, and image data in different formats can be processed, thereby improving the processing ability and accuracy.

[0120] It should be understood that the processor in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0121] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchlink DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0122] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions according to the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that contains one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, or a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0123] It should be understood that the term "and / or" in this document is merely a description of the association relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. In addition, the character " / " in this document generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood by referring to the context before and after.

[0124] In the present invention, "at least one" means one or more, and "a plurality" means two or more. "At least one of the following" or its similar expressions refer to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, and c can be single or multiple.

[0125] It should be understood that in various embodiments of the present invention, the magnitudes of the sequence numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0126] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in connection with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0127] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the devices, apparatuses, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0128] In several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of devices or units can be in electrical, mechanical, or other forms.

[0129] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0130] In addition, the functional units in each embodiment of the present invention can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.

[0131] If a function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or a part of this technical solution can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0132] An embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the data entry method for a Web form based on character recognition as in the method embodiment.

[0133] The computer-readable storage medium provided by the present invention can implement the steps and effects of the data entry method for a Web form based on character recognition in the above method embodiment. To avoid repetition, the present invention will not elaborate further.

[0134] The beneficial effects brought by the technical solution provided by the embodiments of the present invention at least include:

[0135] In the embodiments of the present invention, by obtaining the image data to be entered, constructing a Web form data set according to the image data, and configuring the filling fields and form styles of the Web form according to the Web form data set, it is ensured that the system can accurately fill in the data. According to the filling fields and form styles of the Web form, anchor fields and recognition regions are selected. Further, according to the filling fields and form styles of the Web form, a corresponding Web form text recognition plugin is selected. Through the Web form text recognition plugin, text content is extracted from the recognition region of the image data, and the text content is automatically filled into the corresponding fields of the Web form according to the anchor fields, generating a filled form. The data entry efficiency is greatly improved through automated steps, human errors are reduced, costs are lowered, and image data in different formats can be processed, thereby improving the processing ability and accuracy.

[0136] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

[0137] The following points need to be explained:

[0138] (1) The accompanying drawings of the embodiments of the present invention only relate to the structures involved in the embodiments of the present invention, and other structures can refer to the general design.

[0139] (2) For clarity, in the accompanying drawings used to describe the embodiments of the present invention, the thickness of layers or regions is enlarged or reduced, that is, these drawings are not drawn to actual scale. It can be understood that when an element such as a layer, film, region or substrate is referred to as being "on" or "under" another element, the element can be "directly" on or under the other element or there can be intervening elements.

[0140] (3) Without conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0141] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A data entry method for a Web form based on text recognition, characterized in that: include: S1: Obtain image data to be recorded; S2: constructing a Web table data set according to the image data; S3: configuring the reporting fields and form style of the Web form according to the Web form data set; S4: selecting an anchor field and an identification area according to the reporting fields and form style of the Web form; S5: selecting a corresponding Web form text recognition plug-in according to the reporting fields and form style of the Web form; S6: extracting text content from the recognition area through the Web form text recognition plug-in; S7: Automatically fill the text content into the corresponding field of the Web form according to the anchor field to generate a filling form.

2. The data entry method for a Web form based on text recognition according to claim 1, characterized in that: The image data specifically includes: local image files, screenshot files and scan files.

3. The data entry method for a Web form based on text recognition according to claim 1 is characterized in that: The Web table data set specifically includes: a business field name, a business field type, and a business field length.

4. The data entry method for a Web form based on text recognition according to claim 1, characterized in that: The step of extracting the text content from the recognition area of ​​the image data in S6 specifically includes: S601: Preprocessing the identified area by edge detection algorithm to generate an edge image; S602: extracting a column feature sequence of the edge image through the HOG descriptor; S603: Processing the column feature sequence through a bidirectional LSTM and outputting a feature vector; S604: Extract the text content according to the feature vector in combination with CTC and vocabulary.

5. The method for entering data into a Web form based on text recognition according to claim 4, characterized in that: The S601 specifically includes: S6011: Convert the color image in the recognition area into a grayscale image; S6012: Smoothing the grayscale image using a Gaussian filter; S6013: generating a smoothed image through a convolution operation based on the smoothed grayscale image; S6014: Process the smoothed image using a Canny edge detection algorithm to generate an edge image.

6. The method for entering data into a Web form based on text recognition according to claim 4, characterized in that: The S602 specifically includes: S6021: combining Edge Boxes and ACF detector to identify a text region candidate frame in the edge image; S6022: Classifying the text region candidate boxes by using random forest, and filtering out proposal boxes from the text region candidate boxes; S6023: Perform regression optimization on the proposed frame; S6024: According to the proposal box after regression optimization, the column feature sequence is extracted through the HOG descriptor.

7. The method for entering data into a Web form based on text recognition according to claim 4, characterized in that: The S603 is specifically as follows: According to the following formula, the column feature sequence is processed by a bidirectional LSTM to output a feature vector: Among them, S t represents the memory state at time t, S t-1 represents the memory state at time t-1, f() represents the activation function, W forget represents the weight matrix of the forget gate, C t represents the column feature sequence at time t, W in represents the weight matrix of the input gate, W c represents the weight matrix of the candidate memory unit, Y t represents the eigenvector at time t, W out Represents the weight matrix of the output gate, and g() represents the tanh function.

8. The method for entering data into a Web form based on text recognition according to claim 4, characterized in that: The S604 specifically includes: S6041: Calculate the probabilities of all output paths according to the feature vector through the CTC; S6042: Determine a loss function of the CTC according to the probability; S6043: With the goal of minimizing the loss function, select the best output path in combination with the vocabulary; S6044: Extracting the text content according to the optimal output path.

9. A data entry system for Web forms based on text recognition, characterized in that: include: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the method for entering data into a Web form based on text recognition according to any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for entering data into a Web form based on text recognition as claimed in any one of claims 1 to 8 is implemented.

Images