CN116311300A - Table generating method, device, electronic device and storage medium - Google Patents

Abstract

The disclosure discloses a table generation method, device, electronic equipment and storage medium, and relates to the technical field of computers, in particular to the technical field of optical character recognition. The specific implementation scheme is: acquiring text content information and layout information of the target page, where the layout information of the target page includes page parameter information, cell layout information and text layout information. According to the page parameter information, the cell layout information and the text layout information, the cell coordinate information is obtained. According to the text layout information and the cell coordinate information, the text coordinate information is obtained. According to the cell coordinate information, the text coordinate information and the text content information, a target table and label data corresponding to the target table are generated.

Description

Table generating method, device, electronic device and storage medium

technical field

The present disclosure relates to the technical field of computers, in particular to the technical field of optical character recognition, and in particular to a table generation method, device, electronic equipment and storage medium.

Background technique

Optical character recognition technology refers to the process of analyzing and processing image files after scanning text materials to obtain text and layout information.

With the development of optical character recognition technology, the trained form structured model can be used to identify and extract text data in form images. In the process of training the table structured model, a large amount of sample table images and labeled data of the sample table images are required.

Contents of the invention

The present disclosure provides a table generation method, device, electronic equipment and storage medium.

According to an aspect of the present disclosure, a method for generating a table is provided, including:

The text content information and the layout information of the target page are acquired, and the layout information of the target page includes page parameter information, cell layout information and text layout information. According to the page parameter information, the cell layout information and the text layout information, the cell coordinate information is obtained. According to the text layout information and the cell coordinate information, the text coordinate information is obtained. A target table and label data corresponding to the target table are generated according to the cell coordinate information, the text coordinate information and the text content information.

According to another aspect of the present disclosure, a form generation device is provided, including: a first obtaining module, a first obtaining module, a second obtaining module and a generating module. Wherein, the first obtaining module is used to obtain text content information and layout information of the target page, and the layout information of the target page includes page parameter information, cell layout information and text layout information. The first obtaining module is used to obtain cell coordinate information according to page parameter information, cell layout information and text layout information. The second obtaining module is used to obtain text coordinate information according to text layout information and cell coordinate information. The generating module is used to generate a target table and label data corresponding to the target table according to cell coordinate information, text coordinate information and text content information.

According to another aspect of the present disclosure, there is provided an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; Instructions executed by the at least one processor to enable the at least one processor to perform the method as described above.

According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the above method.

According to another aspect of the present disclosure, there is provided a computer program product comprising a computer program which, when executed by a processor, implements the method as described above.

It should be understood that what is described in this section is not intended to identify key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will be readily understood through the following description.

Description of drawings

The accompanying drawings are used to better understand the present solution, and do not constitute a limitation to the present disclosure. in:

FIG. 1 schematically shows an exemplary system architecture to which a table generating method and device can be applied according to an embodiment of the present disclosure;

Fig. 2 schematically shows a flowchart of a table generation method according to an embodiment of the present disclosure;

Fig. 3 schematically shows a schematic diagram of generating cell coordinate information according to an embodiment of the present disclosure;

Fig. 4 schematically shows a schematic diagram of cells in a table according to an embodiment of the present disclosure;

Fig. 5 schematically shows a schematic diagram of the display effect of text in a cell determined according to the text arrangement according to an embodiment of the present disclosure;

Fig. 6 schematically shows a schematic diagram of a table according to some embodiments of the present disclosure;

Fig. 7 schematically shows a schematic diagram of a table with border lines according to some embodiments of the present disclosure;

Fig. 8 schematically shows a schematic diagram of a form with stamps according to some embodiments of the present disclosure;

Fig. 9 schematically shows a schematic diagram of a filtered table according to some embodiments of the present disclosure;

Fig. 10 schematically shows a block diagram of a table generating device according to an embodiment of the present disclosure; and

Fig. 11 schematically shows a block diagram of an electronic device suitable for implementing a table generation method according to an embodiment of the present disclosure.

Detailed ways

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and they should be regarded as exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

In the process of using optical recognition technology to identify and extract information from form images, it is not only necessary to recognize the text in the form image, but also to identify the structure of the form, and perform structured processing of the row and column information on the recognition results.

Therefore, before training the model used to recognize the text and table structure of the table image, it is necessary to collect a large number of sample table images, and label the text and table structure in the sample table image to obtain the labeled data. Using the sample table image and the labeled data Train the model.

However, in related technologies, manual labeling is used to obtain labeled data, which not only has low labeling efficiency but also low accuracy of labeling results, especially for table images in the financial field, such as: balance sheet, cash flow statement, Income statement, etc., the table structure in these table images is relatively complex, and there are great difficulties in labeling the data in the table image by manual labeling. It directly affects the effect of model training.

In view of this, an embodiment of the present disclosure provides a table generation method, including:

The text content information and the layout information of the target page are acquired, and the layout information of the target page includes page parameter information, cell layout information and text layout information. According to the page parameter information, the cell layout information and the text layout information, the cell coordinate information is obtained. According to the text layout information and the cell coordinate information, the text coordinate information is obtained. A target table and label data corresponding to the target table are generated according to the cell coordinate information, the text coordinate information and the text content information. It can generate tables with complex structures applied in the field of financial technology, and automatically generate label data corresponding to the tables while generating the tables, eliminating the need for manual labeling.

Fig. 1 schematically shows an exemplary system architecture to which the method and device for generating a table can be applied according to an embodiment of the present disclosure.

It should be noted that, what is shown in FIG. 1 is only an example of the system architecture to which the embodiments of the present disclosure can be applied, so as to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other device, system, environment or scenario. For example, in another embodiment, the exemplary system architecture to which the form generation method and apparatus can be applied may include terminal equipment, but the terminal equipment can implement the form generation method and apparatus provided by the embodiments of the present disclosure without interacting with the server .

As shown in FIG. 1 , a system architecture 100 according to this embodiment may include terminal devices 101 , 102 , 103 , a network 104 and a server 105 . The network 104 is used as a medium for providing communication links between the terminal devices 101 , 102 , 103 and the server 105 . Network 104 may include various connection types, such as wired and/or wireless communication links, among others.

Users can use terminal devices 101 , 102 , 103 to interact with server 105 via network 104 to receive or send messages and the like. Various communication client applications can be installed on the terminal devices 101, 102, 103, such as knowledge reading applications, web browser applications, search applications, instant messaging tools, email clients and/or social platform software, etc. (only example).

The terminal devices 101, 102, 103 may be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers and the like.

The server 105 may be a server that provides various services, such as a background management server that supports content browsed by users using the terminal devices 101 , 102 , 103 (just an example). The background management server can analyze and process received data such as user requests, and feed back processing results (such as webpages, information, or data obtained or generated according to user requests) to the terminal device.

It should be noted that, generally, the method for generating a table provided by the embodiment of the present disclosure may be executed by the terminal device 101 , 102 , or 103 . Correspondingly, the table generation apparatus provided by the embodiment of the present disclosure may also be set in the terminal device 101 , 102 , or 103 .

Alternatively, the table generation method provided by the embodiment of the present disclosure may generally be executed by the server 105 . Correspondingly, the table generating apparatus provided by the embodiment of the present disclosure can generally be set in the server 105 . The form generation method provided by the embodiments of the present disclosure may also be executed by a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101 , 102 , 103 and/or the server 105 . Correspondingly, the form generation apparatus provided by the embodiments of the present disclosure may also be set in a server or a server cluster that is different from the server 105 and can communicate with the terminal devices 101 , 102 , 103 and/or the server 105 .

For example, when a user sends a form generation request online, the terminal devices 101, 102, and 103 can obtain the text content information and the layout information of the target page selected by the user from the database, and then send the acquired text content information and the layout information of the target page To the server 105, the server 105 obtains the cell coordinate information according to the page parameter information, cell layout information and text layout information. According to the text layout information and the cell coordinate information, the text coordinate information is obtained. And according to the cell coordinate information, text coordinate information and text content information, generate a target table and label data corresponding to the target table. Alternatively, a server or server cluster capable of communicating with the terminal devices 101, 102, 103 and/or the server 105 analyzes the target content, and finally extracts the content that the user is interested in.

It should be understood that the numbers of terminal devices, networks and servers in Fig. 1 are only illustrative. According to the implementation needs, there can be any number of terminal devices, networks and servers.

Fig. 2 schematically shows a flowchart of a table generation method according to an embodiment of the present disclosure.

As shown in FIG. 2, the method includes operations S210-S240.

In operation S210, the text content information and the layout information of the target page are acquired, and the layout information of the target page includes page parameter information, cell layout information and text layout information.

In operation S220, cell coordinate information is obtained according to the page parameter information, cell layout information, and text layout information.

In operation S230, the text coordinate information is obtained according to the text layout information and the cell coordinate information.

In operation S240, a target table and label data corresponding to the target table are generated according to the cell coordinate information, text coordinate information, and text content information.

According to an embodiment of the present disclosure, the text content information may be table content information for displaying in a cell of a table. For example: in the XX transaction list form, the table content information can include: account information, account name information, query start date information, query deadline information, query time information, query teller information, currency information, transaction date information, record Account date, transaction location information, transaction type information, loan status information, transaction amount information, etc.

According to an embodiment of the present disclosure, the above text content information may be configured according to the content information of the real form, and the configured text content information may be stored in the table header dictionary as a table content candidate set. In order to obtain the required text content information by traversing the header dictionary.

According to an embodiment of the present disclosure, the page parameter information may include page width information, page length information, page row number information, and page column number information.

According to an embodiment of the present disclosure, the cell layout information may include column width ratio information and relative position information between cells. The column width ratio information can generally be configured for the row with the largest number of columns in the table, based on the smallest cell, and the width of other cells is scaled based on the width of the smallest cell according to the above-mentioned column width ratio information. For example: 3 cells can be included in the same row, and the column width ratio between the above 3 cells is as follows: the column width ratio of the first cell to the smallest cell in the row is "2", the second The column width ratio of the cell to the smallest cell in the row is "1.7", and the third cell is the smallest cell in the row.

For example: the page width is 100, the page length is 300, the number of page rows is 10, and the row with the largest number of page columns includes 4 columns of cells. The column width ratio of the column from left to right can be: the column width ratio of the first cell is "1", the column width ratio of the second cell is "1.5", and the column width of the third cell If the ratio is "1" and the column width ratio of the fourth cell is "1.5", then the width of the cells in this column can be obtained in order: 20, 30, 20, 30.

According to the embodiments of the present disclosure, similarly, the height information of the smallest cell in a certain column may also be obtained according to the page length and page row height ratio information. The height of other cells is scaled according to the row height ratio to obtain the height information of all cells in a column, which will not be described here.

According to an embodiment of the present disclosure, the relative position information between cells may represent relative position information between cells in a certain row and cells in other rows. For example: the first row may include 4 cells, which are cell CellA ₁ , cell CellA ₂ , cell CellA ₃ , and cell CellA ₄ . The second row may include 7 cells, which are cell CellB ₁ , cell CellB ₂ , cell CellB ₃ , cell CellB ₄ , cell CellB ₅ , cell CellB ₆ , and cell CellB ₇ . The relative position relationship between cells can be expressed as: cell CellA ₁ (1, 1), cell CellA ₂ (2, 3), cell CellA ₃ (4, 5), cell CellA ₄ (6, 7 ).

According to the embodiment of the present disclosure, according to the relative positional relationship between the cells, it can be known that the cell CellA ₁ and the cell CellB ₁ are in a vertically aligned positional relationship, that is, the cell CellA ₁ and the cell CellB ₁ have the same width. Cell CellA ₂ spans cell CellB ₂ and cell CellB ₃ , that is, the width of cell CellA ₂ is equal to the sum of the widths of cell CellB ₂ and cell CellB ₃ . And so on, so I won't repeat them here.

For example: the width of the cells in the second row can be: the width of the cell CellB ₁ is "10", the width of the cell CellB ₂ is "15", the width of the cell CellB ₃ is "10", the width of the cell CellB ₄ has a width of "5", cell CellB ₅ has a width of "20", cell CellB ₆ has a width of "20", and cell CellB ₇ has a width of "20". According to the relative positional relationship between the above cells, the width of the cell CellA ₁ is "10", the width of the cell CellA ₂ is "25", the width of the cell CellA ₃ is "25", and the width of the cell CellA _{4 is} has a width of "40".

According to an embodiment of the present disclosure, the text layout information may include an arrangement manner of text in a cell, and font size information of the text. The way of arranging the text in the cell may include aligning the text at the top of the cell, aligning the center, and aligning the bottom of the cell. The font size information of the text may include information about the height of the text, information about the width of the text, and the like.

According to an embodiment of the present disclosure, taking the top-aligned arrangement of the text in the cell as an example, and the text is arranged in a single row in the cell, it can be determined that the height of the text is equal to the height of the cell.

For example: the width of the first cell CellA ₁ in the first row is "10", the height of the text can be "5", and the coordinate information of the cell CellA ₁ can be determined: "(0,0), (10,0 ), (0,5), (10,5)".

According to an embodiment of the present disclosure, it is assumed that texts are arranged in a top-aligned manner in a cell as an example, and the texts are arranged in a single row in a cell. For example: the width of the text can be "1", the text content information in the first cell CellA ₁ in the first row can be "account number", then the coordinates of the word "account" can be "(0 ,0), (1,0), (0,5), (1,5)". The coordinates of the word "number" can be "(1,0), (2,0), (1,5), (2,5)".

According to the embodiment of the present disclosure, the text content can be filled into the corresponding cell according to the text coordinate information, and the target table can be obtained.

According to an embodiment of the present disclosure, the annotation data corresponding to the target table represents structured annotation data combined with text content information and cell coordinate information. For example, the label data of the cell CellA ₁ may include: the coordinate information of the cell CellA ₁ and the text content information of the cell CellA ₁ .

According to an embodiment of the present disclosure, during the process of generating the target table, the text content information may include header information and random number information corresponding to the header. For example: the header information can be "account number", and the random number information corresponding to the header can be "000XXX1111".

According to the embodiments of the present disclosure, target tables with different cell layouts and text layouts can be flexibly generated by acquiring text content information and target page layout information, realizing rapid generation of target tables for model training and corresponding target tables. Annotate the data. Since the labeling data is the cell coordinate information, text coordinate information and text content information obtained directly in the process of generating the target table, the accuracy and efficiency of the labeling data are improved.

The method shown in FIG. 2 will be further described below in conjunction with specific embodiments with reference to FIGS. 3 to 9 .

Fig. 3 schematically shows a schematic diagram of generating cell coordinate information according to an embodiment of the present disclosure.

As shown in FIG. 3 , in step 300 , minimum cell size information 3203 is obtained according to page parameter information 3201 and cell size ratio information 3202 . Then the cell size information 3205 is obtained according to the minimum cell size information 3203 and the relative position relationship information 3204 . According to the cell size information 3205 and the page parameter information 3201, the arrangement position information 3206 of the cells on the target page is obtained. Multiple target cells 3207 in the same row can be determined according to the arrangement position information 3206 of the cells on the target page. According to the newline identification information 3208, the information 3209 on the number of lines arranged in the cell can be obtained. The height information 3210 of a plurality of target cells can be obtained according to the information 3209 on the number of rows of characters arranged in the cell. Cell coordinate information 3211 can be obtained according to the multiple target cells 3207 located in the same row and the height information 3210 of the multiple target cells.

According to an embodiment of the present disclosure, the size ratio information of a cell may include column width ratio information and row height ratio information. The following takes the column width ratio information as an example to describe the process of determining the cell width information in detail.

According to the embodiments of the present disclosure, generally, the row with the largest number of columns in the target table is selected for setting the column width ratio information of the cell. For example: the page width is 100, the row with the most columns can include 3 columns of cells, and the column width ratio information is as follows: the column width ratio of the cells in the first column can be "6", and the column width ratio of the cells in the second column It can be "1", the column width ratio of the third column cell can be "3", and the minimum cell width can be 10.

According to an embodiment of the present disclosure, the relative positional relationship between the cells may represent the column width span information between the column where the cell is located and the columns where other cells are located. For example: the cell CellA ₁ can be the cell located in the first row and the first column, and the cells CellB ₁ and CellB ₂ can be the cells located in the second row, the first column and the second column respectively. The column width of the column where cell CellA ₁ is located spans the column width of the column where cell CellB ₁ and cell CellB ₂ are located, then the relative positional relationship between the cells can be expressed as: cell CellA ₁ (cell CellB ₁ , cell CellB ₂ ). That is, the width of cell CellA ₁ is equal to the sum of the widths of cell CellB ₁ and cell CellB ₂ .

According to the minimum cell width and the relative positional relationship between cells, the cell width information can be obtained according to formula (1):

Cell width = minimum cell width × SUM [column width ratio (column width span information)] (1)

For example: the minimum cell width is 10, the column width ratio of cell CellB ₁ is 2, the column width ratio of cell CellB ₂ is 1.5, and the column width span information is: cell CellA ₁ (cell CellB ₁ , cell CellB ₂ ), the width of cell CellA ₁ can be obtained as 35.

According to an embodiment of the present disclosure, the text layout information includes the line break identification information of the text in the cell, and the cell coordinate information is obtained according to the cell width information, page parameter information, and text layout information, which may include the following operations:

Identify the newline identification information, and get the information on the number of rows of text in the cell. According to the cell width information and the page parameter information, the arrangement position information of the cells on the target page is obtained. Cell coordinate information is obtained according to the arrangement position information and the arrangement row number information.

According to the embodiment of the present disclosure, in a real table, the text in each cell may be arranged across rows. Therefore, by setting the line break identification information of the text in the cell, the text in the cell can be obtained. Arrange the number of rows information.

For example: in the cell CellA ₁ , the text content information may be "account/card number:\nAccount/Card.No:", where "\n" may indicate newline identification information.

According to an embodiment of the present disclosure, by identifying the newline identification information, it can be obtained that the number of rows of the text content information in the cell CellA ₁ is 2 rows.

According to the embodiment of the present disclosure, since the minimum cell width is set with the number of columns at most one row, the width of other cells is also obtained according to the column width span information between the cell and the cell on the row where the minimum cell is located . Therefore, the arrangement position information of the cells on the target page can be obtained according to the cell width information and the page parameter information.

For example, the row with the largest number of columns may be the sixth row, and the column width and span information of other cells may be determined based on the relative positional relationship between the cells in the sixth row. For example: the column width span information of the first cell in the fifth row can be (1, 2), which means that the column width of the first cell in the fifth row is equal to the first cell and the first cell in the sixth row The sum of the widths of the 2 cells. The column width span information of the first cell in row 4 can be (1, 2, 3), which means that the column width of the first cell in row 4 is equal to the first cell in row 6～ The sum of the width of the 3rd cell.

According to the embodiments of the present disclosure, since the current height of each cell only considers the page parameter information and the row height ratio, in a real table image, the heights of the cells in each row should be the same. Therefore, on the basis of determining the relative arrangement position of the cells on the target page, the height information of the cells in the row can be finally determined by combining the heights of multiple cells in the same row.

According to an embodiment of the present disclosure, obtaining the cell coordinate information according to the arrangement position information and the arrangement row number information may include the following operations:

According to the arrangement position information, multiple target cells in the same row are obtained. According to the arrangement row number information, the height information of multiple target cells is obtained. According to the arrangement position information and the height information, the cell coordinate information is obtained.

For example: according to the arrangement position information, multiple target cells located in the same row can be determined, for example: the target cell can include 3 cells, which are cell CellM ₁ , cell CellM ₂ , and cell CellM ₃ . The height information of each cell can be obtained according to the number of rows of text in each cell, for example: the number of rows of text in cell CellM ₁ is 2 rows, and the number of rows of text in cell CellM ₂ The number is 3 rows, and the number of rows of text in cell CellM ₃ is 2 rows.

According to an embodiment of the present disclosure, obtaining the height information of a plurality of target cells according to the number of arranged rows may include the following operations:

Sorting the row number information of multiple target cells to obtain a sorting result. According to the sorting results, the height information is obtained.

For example, the above 3 cells are sorted according to the row number information, and the sorting result can be: cell CellM ₂ , cell CellM ₁ , and cell CellM ₃ . In each row, the height of the cell with the largest number of character arrangement rows can be determined as the final height of all the cells in the row. For example: the height of the cell CellM ₂ can be 5, then the cell CellM ₁ and the cell CellM ₃ are also 5.

Fig. 4 schematically shows a schematic diagram of cells in a table according to an embodiment of the present disclosure.

As shown in FIG. 4 , in 400 , cells A to P can all represent the cells where the head of the cross-column is located. Among them, the row with the largest number of columns may be the row where the cell Q is located, and this row represents the table header. The column width of cell A spans the width of 6 cells from cell Q to cell V. Cell B, Cell F, Cell J, Cell N have the same column width as Cell Q. The column width of cells C, G, K, and O is equal to the sum of the column widths of cells R and S. The column width of cell D, cell H, and cell L is equal to the sum of the column widths of cell T and cell U. The column width of cell P is equal to the sum of the column widths of cell T, cell U, and cell V. The column width of cell E, cell I, and cell M is equal to the column width of cell V. From top to bottom, from the row where cell a1 is located to the row where cell an is located, they are all cells where the body of the table is located, and the structure of the cell where the table body is located is the same as that of the cell where the header is located.

According to an embodiment of the present disclosure, the text layout information includes text layout information in the cell, text size information, and text line break identification information in the cell. The above operation S230 may include the following operations:

Determine the target strategy based on the arrangement information. According to the newline identification information, the information on the number of rows of the text in the cell is obtained. Based on the target strategy, the text coordinate information is obtained according to the cell coordinate information, text size information, arrangement mode information and arrangement row number information.

According to an embodiment of the present disclosure, the arrangement of characters may include a horizontal arrangement and a vertical arrangement of the characters in the cell. Wherein, the arrangement in the horizontal direction may include: left alignment, center alignment, and right alignment. The vertical arrangement methods may include: top alignment, center alignment and bottom alignment. The arrangement of text in the cell can include the following 9 ways: Left Alignment + Top Alignment, Left Alignment + Center Alignment, Left Alignment + Bottom Alignment, Center Alignment + Top Alignment, Center Alignment + Center Alignment, Center Align + Bottom, Right + Top, Right + Center, Right + Bottom.

Fig. 5 schematically shows a schematic diagram of a display effect of text in a cell determined according to a text arrangement method according to an embodiment of the present disclosure.

As shown in Figure 5, the display effect of left alignment + top alignment is shown in 5321, and the text is arranged from the top left corner of the cell. The display effect of center alignment + center alignment is shown in 532i, and the text is arranged from the middle of the cell. The display effect of right alignment + bottom alignment is shown in 532I. The text is arranged from the bottom right of the cell, and the last text is located at the bottom right top corner of the cell.

According to an embodiment of the present disclosure, the target strategy may represent a calculation strategy for different text coordinates configured in different arrangements.

According to an embodiment of the present disclosure, the text size information includes text height information and text width information. Based on the target strategy, the text coordinate information can be obtained according to the cell coordinate information, text size information, layout information, and layout row number information, which can include the following operations:

According to the cell coordinate information, the text width information and the layout information, the text abscissa information is obtained. Based on the target strategy, the vertical coordinate information of the text is obtained according to the cell coordinate information, the text height information and the arrangement row number information. The text coordinate information is obtained according to the text abscissa information and the text ordinate information.

According to an embodiment of the present disclosure, a custom offset parameter threshold range can be configured according to the horizontal arrangement of text in the cell, for example: [0, 1]. When the horizontal arrangement of the text in the cell is aligned to the left, it can be determined that the offset parameter of the text in the horizontal direction is 0. When the text is aligned in the center in the horizontal direction in the cell, the offset parameter of the text in the horizontal direction can be determined to be 0.5. When the text is aligned to the right in the horizontal direction in the cell, the offset parameter of the text in the horizontal direction can be determined to be 1.

According to the embodiment of the present disclosure, in a real table, the border line of the table will occupy a certain width. In order to prevent the text close to the cell border from being partially covered, the threshold range of the offset parameter above can be changed according to actual needs. For example: can be changed to [0.2, 0.8].

According to an embodiment of the present disclosure, obtaining the text abscissa information according to the offset parameter information, text width information, and cell coordinate information may include the following operations:

According to the coordinate information of the left vertex and the coordinate information of the right vertex, the cell width information is obtained. In the case that the arrangement information is determined to be center-aligned, the abscissa information of the text is obtained according to the left vertex coordinate information, the cell width information, and the text width information. In the case where it is determined that the arrangement information is left-aligned, the text abscissa information is obtained according to the coordinate information of the left vertex and the offset parameter information. When it is determined that the arrangement information is right-aligned, the abscissa information of the text is obtained according to the coordinate information of the right vertex, the text width information, and the offset parameter information.

According to an embodiment of the present disclosure, the text width information may represent the width information of all texts in the cell obtained according to a preset text size.

According to an embodiment of the present disclosure, when it is determined that the arrangement information is centered, the abscissa information of the text can be obtained according to the formula (2):

Text abscissa = left vertex coordinate + cell width/2-text width/2(2)

For example: the abscissa of the left vertex of the cell can be 0, the cell width can be 5, the text width can be 3, and the initial abscissa of the text can be determined as 1.

According to an embodiment of the present disclosure, when it is determined that the arrangement information is left-aligned, the text abscissa information can be obtained according to formula (3):

Text abscissa = left vertex coordinate + offset parameter (3)

For example: the abscissa of the left vertex of the cell can be 0, the offset parameter corresponding to the left-aligned arrangement can be 0.2, and the starting abscissa of the text can be determined to be 0.2.

According to an embodiment of the present disclosure, when it is determined that the arrangement information is right-aligned, the text abscissa information can be obtained according to formula (4):

Text abscissa=right vertex coordinate-text width-offset parameter (4)

For example: the abscissa of the right vertex of the cell can be 5, the text width can be 3, the offset parameter corresponding to the right-aligned arrangement can be 0.8, and the starting abscissa of the text can be determined to be 1.2.

According to the embodiment of the present disclosure, in the case of determining the initial coordinates of the text in the cell, the abscissa of the next text can be translated to the right along the horizontal direction by the width of one text font size on the basis of the initial coordinates, that is The abscissa information of the next text can be obtained.

According to an embodiment of the present disclosure, the cell coordinate information may include the coordinate information of the upper left vertex of the cell. Based on the target strategy, the text vertical coordinate information may be obtained according to the cell coordinate information, text height information, and row number information, which may include the following operate:

According to the cell coordinate information, the minimum cell height information is obtained. According to the coordinate information of the upper left vertex, the minimum cell height information, the text height information and the arrangement row number information, the text vertical coordinate information is obtained.

For example: when the coordinate information of each cell is determined, the height information of the smallest cell on the target page can be obtained, for example: the height of the smallest cell can be 5. The height of the text can be 1, and the number of rows can include the information of the number of rows of the text in the cell and the information of the maximum number of rows of the text in multiple cells in the same row. For example: the number of rows of text in a cell can be 1, and the maximum number of rows of text in multiple cells in the same row can be 2. The vertical coordinate information of the text can be obtained according to different arrangements of the text in the vertical direction of the cell and according to different calculation strategies.

According to an embodiment of the present disclosure, in the case where the arrangement information is determined to be top-aligned, the vertical coordinate information of the text can be obtained according to the formula (5):

Text vertical coordinate = vertical coordinate of upper left vertex + minimum cell height × maximum number of rows/2-text height×maximum number of rows/2 (6)

For example: the ordinate of the upper left vertex can be 0, the minimum cell height is 5, the maximum number of rows is 2, the text height is 1, and the text ordinate can be determined to be 4.

According to an embodiment of the present disclosure, in the case where it is determined that the arrangement information is centered, the vertical coordinate information of the text can be obtained according to the formula (6):

Text ordinate = vertical coordinate of the upper left vertex + minimum cell height × maximum number of rows/2-text height × number of rows of text in this cell/2 (7)

For example: the ordinate of the upper left vertex can be 0, the minimum cell height is 5, the maximum number of rows is 2, the height of the text is 1, the number of rows of text in this cell is 1, and the ordinate of the text can be determined as 4.5.

According to an embodiment of the present disclosure, when it is determined that the arrangement information is bottom-aligned, the vertical coordinate information of the text can be obtained according to the formula (7):

Text ordinate = vertical coordinate of upper left vertex + minimum cell height × maximum number of rows/2-text height×maximum number of rows/2+minimum cell height×(maximum number of rows-in this cell The number of lines of text layout) (7)

For example: the ordinate of the upper left vertex can be 0, the minimum cell height is 5, the maximum number of rows is 2, the height of the text is 1, the number of rows of text in this cell is 1, and the ordinate of the text can be determined as 9.

According to an embodiment of the present disclosure, the above operation S240 may include the following operations:

Determine the target cell according to the text content information. Fill the text content information into the target cell according to the text coordinate information. Arrange the target cells on the target page according to the cell coordinate information to generate the target table. The cell coordinate information, text coordinate information and text content information are processed according to a predetermined data format to obtain label data.

According to an embodiment of the present disclosure, the annotation data may include table content information and table structure information. For example: the text content information can be "account number", and the target cell can be determined to be the first cell in the second row. According to the text coordinate information, for example: "Account" is filled in the target cell according to the starting abscissa of the first text of 0.2 and the starting ordinate of 1. "Number" is filled in the target cell according to the starting abscissa of the second text, which is 0.5, and the starting ordinate, 1. And the target cells are arranged on the target page according to the cell coordinate information to obtain the target table.

It should be noted that, in the embodiment of the present disclosure, the coordinate information of the cell and the coordinate information of the text may both be pixel coordinate information.

According to an embodiment of the present disclosure, the above-mentioned cell coordinate information, text coordinate information, and text content information can be stored according to a one-to-one mapping relationship, and processed according to a predetermined data format, for example, it can be in JSON format to obtain label data .

Fig. 6 schematically shows a table diagram according to some embodiments of the present disclosure.

As shown in FIG. 6 , in 600 , in the cells in the first row, the arrangement of the text in the cells is center alignment in the horizontal direction + center alignment in the vertical direction. In each cell after row 2, the arrangement of the text in the cell is left alignment in the horizontal direction + center alignment in the vertical direction. In the cells of the 2nd row, the 4th row to the 5th row, the number of rows in which the text is arranged in the cell is 2 rows. In the first cell of the third row, the number of rows of text in the cell is 3 rows.

It should be noted that since the form generated in the embodiment of the present disclosure is used as a sample form for model training, the text corresponding to "account number", "card number", "transaction amount" and "transaction date" in the content of the form The values are random numbers.

In order to make the target table closer to the collected table image, different table border lines can be added to the target table.

According to an embodiment of the present disclosure, the above table generation method may further include the following operations:

Get the table border line style template. According to the table border line template and according to the cell coordinate information, a border line is constructed on the target table to obtain the first table image.

According to an embodiment of the present disclosure, the target page may be a blank page constructed using Pillow image processing software. It is also possible to randomly add grayscale colors on a blank page to obtain first form images of different grayscales.

According to an embodiment of the present disclosure, the table border line template may include multiple line styles, for example: full-frame line style, non-full-frame line style, dashed line style, and the like. According to actual needs, the linear in the table border line template can be flexibly selected to construct border lines for the cells in the target table to obtain the first table image.

Fig. 7 schematically shows a schematic diagram of a table with border lines according to some embodiments of the present disclosure.

As shown in FIG. 7 , in 700 , among the cells in the first row to the sixth row, the border line type of the cell is full-frame linear. The cells from the seventh row to the last row include various border line styles. For example: in the first cell in row 7, the line type of the left border and top border is a solid border line type, and the line type of the right border and bottom border is no border line. In the second cell of row 7, the line type of the left border and bottom border is no border line, the line type of the top border is solid border line type, and the line type of the right border is dashed border line type.

According to an embodiment of the present disclosure, the above table generation method may further include the following operations:

Get stamp material image set. Randomly select the target stamp image from the stamp material image set. The target stamp image and the first form image are processed to obtain the second form image.

According to an embodiment of the present disclosure, the stamp material image set may include stamp template images simulating different scenes in real form images. The target stamp image T can be randomly selected from the stamp template images, and the target stamp image T can be fused with the first form image to obtain the second form image.

Fig. 8 schematically shows a schematic diagram of a form with stamps according to some embodiments of the present disclosure.

As shown in Figure 8, in 1100, the target seal image T is the seal of XX company, the position of the target seal image T can be random, in other words, change the relative position of the target seal image T and the first form image, can get many A second table image is used as a sample table image for model training.

In the collected real form image, the seal may be incomplete, for example: for the seam seal, only partial seals are displayed on the form images of different pages. Therefore, in order to increase the closeness between the second form image and the real form image, the target stamp image may be processed first, and then fused with the first form image.

According to an embodiment of the present disclosure, processing the target stamp image and the first form image to obtain the second form image may include the following operations:

The target stamp image is cropped to obtain a partial stamp image. Delete the background color in the partial stamp image to obtain the target partial stamp image. The target partial seal image and the first form image are processed to obtain the second form image.

Since real form images can be collected by image acquisition equipment under different light conditions, when different light conditions, different acquisition methods, and different acquisition equipment collect the image information of the same form, the obtained form images can be different. of. In order to simulate the form images collected under different acquisition environments, the target partial seal image and the first form image are processed to obtain the second form image, which may include the following operations:

The target partial seal image is fused with the first form image to obtain a form image with the target partial seal. Filtering is performed on the form image with the target partial seal to obtain a second form image.

According to an embodiment of the present disclosure, the filtering process may include at least one of the following: Gaussian filtering, contour filtering, detail filtering, edge enhancement filtering, smoothing filtering, depth smoothing filtering, unsharp mask filtering, and the like. The form image with the target partial seal can be filtered one or more times by one or more filtering methods to obtain a second form image that can be used to simulate different acquisition environments.

Fig. 9 schematically shows a schematic diagram of a filtered table according to some embodiments of the present disclosure.

As shown in FIG. 9 , at 900 , it may be a form image for simulating a real acquisition environment obtained by processing a form image with a target seal by using Gaussian filtering. It can be seen from 900 that the fonts in the table image are in a relatively fuzzy state, and using the filtered table image to train the table recognition model can effectively improve the training accuracy of the model.

Fig. 10 schematically shows a block diagram of a table generating device according to an embodiment of the present disclosure.

As shown in FIG. 10 , the table generating apparatus 1000 of this embodiment may include a first obtaining module 1010 , a first obtaining module 1020 , a second obtaining module 1030 and a generating module 1040 .

The first acquisition module 1010 is configured to acquire text content information and layout information of the target page, where the layout information of the target page includes page parameter information, cell layout information and text layout information. In some embodiments, the first obtaining module 1010 may be configured to implement the operation S210 described above, which will not be described in detail here.

The first obtaining module 1020 is configured to obtain cell coordinate information according to page parameter information, cell layout information and text layout information. In some embodiments, the first obtaining module 1020 may be configured to implement operation S220 described above, which will not be repeated here.

The second obtaining module 1030 is configured to obtain text coordinate information according to text layout information and cell coordinate information. In some embodiments, the second obtaining module 1030 may be configured to implement operation S230 described above, which will not be repeated here.

The generating module 1040 is configured to generate the target table and the label numbers corresponding to the target table according to the cell coordinate information, text coordinate information and text content information. In some embodiments, the generating module 1040 may be used to implement the operation S240 described above, which will not be repeated here.

According to an embodiment of the present disclosure, the cell layout information includes size ratio information of the cells and relative positional relationship information between the cells. The first obtaining module may include: a first obtaining submodule, a second obtaining submodule and a third obtaining submodule. Wherein, the first obtaining sub-module is used to obtain the minimum cell size information according to the page parameter information and the cell size ratio information. The second obtaining sub-module is used to obtain cell size information according to the minimum cell size information and relative position relationship information. The third obtaining sub-module is used to obtain cell coordinate information according to cell size information, page parameter information and text layout information.

According to an embodiment of the present disclosure, the text layout information includes line break identification information of the text in the cell, and the third obtaining submodule may include: an identification unit, a first obtaining unit, and a second obtaining unit. Wherein, the recognition unit is used to recognize the newline identification information, and obtain the information on the number of rows of characters arranged in the cell. The first obtaining unit is configured to obtain the arrangement position information of the cells on the target page according to the cell size information and the page parameter information. The second obtaining unit is configured to obtain cell coordinate information according to the arrangement position information and the arrangement row number information.

According to an embodiment of the present disclosure, the second obtaining unit may include: a first obtaining subunit, a second obtaining subunit, and a third obtaining subunit. Wherein, the first obtaining subunit is configured to obtain multiple target cells located in the same row according to the arrangement position information. The second obtaining subunit is used to obtain the height information of multiple target cells according to the arrangement row number information. The third is to obtain subunits, which are used to obtain cell coordinate information according to the arrangement position information and height information.

According to an embodiment of the present disclosure, the second obtaining subunit is configured to sort the row number information of the plurality of target cells to obtain a sorting result. According to the sorting results, the height information is obtained.

According to an embodiment of the present disclosure, the text layout information includes text layout information in a cell, text size information, and text line break identification information in a cell, and the second obtaining module may include: a first determination sub-module, a second Four obtained sub-modules and fifth obtained sub-modules. Wherein, the first determination sub-module is configured to determine the target strategy according to the arrangement information. The fourth obtaining sub-module is used to obtain information on the number of rows of text arranged in the cell according to the newline identification information. The fifth obtaining sub-module is used to obtain the text coordinate information based on the target strategy according to the cell coordinate information, text size information, arrangement mode information and arrangement row number information.

According to an embodiment of the present disclosure, the character size information includes character height information and character width information, and the fifth obtaining submodule may include: a third obtaining unit, a fourth obtaining unit, and a fifth obtaining unit. Wherein, the third obtaining unit is configured to obtain text abscissa information according to cell coordinate information, text width information, and layout information. The fourth obtaining unit is used to obtain text ordinate information based on the target strategy, according to cell coordinate information, text height information, and arrangement row number information. The fifth obtaining unit is configured to obtain text coordinate information according to text abscissa information and text ordinate information.

According to an embodiment of the present disclosure, the third obtaining unit may include a determining subunit and a fourth obtaining subunit. Wherein, the determination sub-unit is used to determine the offset parameter information of the characters in the horizontal direction according to the arrangement information. The fourth obtaining subunit is used to obtain text abscissa information according to the offset parameter information, text width information and cell coordinate information.

According to an embodiment of the present disclosure, the cell coordinate information includes left vertex coordinate information and right vertex coordinate information, and the fourth obtaining subunit is configured to obtain cell width information according to the left vertex coordinate information and the right vertex coordinate information. In the case that the arrangement information is determined to be center-aligned, the abscissa information of the text is obtained according to the left vertex coordinate information, the cell width information, and the text width information. In the case where it is determined that the arrangement information is left-aligned, the text abscissa information is obtained according to the coordinate information of the left vertex and the offset parameter information. When it is determined that the arrangement information is right-aligned, the abscissa information of the text is obtained according to the coordinate information of the right vertex, the text width information, and the offset parameter information.

According to an embodiment of the present disclosure, the fourth obtaining unit may include a fifth obtaining subunit and a sixth obtaining subunit. Wherein, the fifth obtaining subunit is used to obtain the minimum cell height information according to the cell coordinate information. The sixth obtaining subunit is used to obtain text vertical coordinate information according to the upper left vertex coordinate information, minimum cell height information, text height information and arrangement row number information.

According to an embodiment of the present disclosure, the information on the number of rows arranged in a cell includes the information on the number of rows arranged in a cell and the information on the number of rows arranged in a plurality of cells in the same row; the sixth obtaining subunit is used for: In the case that the arrangement mode information is determined to be top-aligned, the longitudinal coordinate information of the text is obtained according to the coordinate information of the upper left vertex, the minimum cell height information, the text height information and the maximum arrangement row number information. In the case of determining whether the arrangement information is center-aligned or bottom-aligned, according to the coordinate information of the upper left vertex, the minimum cell height information, the text height information, the number of rows of text in the cell, and the maximum number of rows , to get the text ordinate information.

According to an embodiment of the present disclosure, the generating module may include a second determining submodule, a filling submodule, a generating submodule and a sixth obtaining submodule. Wherein, the second determining submodule is used to determine the target cell according to the text content information. The filling sub-module is used to fill the text content information into the target cell according to the text coordinate information. The generation sub-module is used to arrange the target cells on the target page according to the cell coordinate information to generate the target table. The sixth obtaining sub-module is used to process cell coordinate information, text coordinate information and text content information according to a predetermined data format to obtain label data.

According to an embodiment of the present disclosure, the above table generation apparatus may further include a second obtaining module and a third obtaining module. Wherein, the second obtaining module is used to obtain the form border line template. The third obtaining module is configured to construct border lines on the target table according to the table border line template and cell coordinate information to obtain the first table image.

According to an embodiment of the present disclosure, the above table generating apparatus may further include a third obtaining module, a selecting module and a fourth obtaining module. Wherein, the third obtaining module is used to obtain the seal material image set. The selection module is used to randomly select the target stamp image from the stamp material image set. The fourth obtaining module is configured to process the target seal image and the first form image to obtain the second form image.

According to an embodiment of the present disclosure, the fourth obtaining module may include a seventh obtaining submodule, an eighth obtaining submodule, and a ninth obtaining submodule. Wherein, the seventh obtaining sub-module is used to perform cropping processing on the target seal image to obtain a partial seal image. The eighth obtaining sub-module is used to delete the background color in the partial seal image to obtain the target partial seal image. The ninth obtaining sub-module is used to process the target partial seal image and the first form image to obtain the second form image.

According to an embodiment of the present disclosure, the ninth obtaining submodule may include a sixth obtaining unit and a seventh obtaining unit. Wherein, the sixth obtaining unit is configured to fuse the target partial seal image and the first form image to obtain the form image with the target partial seal. The seventh obtaining unit is configured to filter the form image with the target partial seal to obtain the second form image.

According to the embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

According to an embodiment of the present disclosure, an electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by at least one processor, and the instructions are processed by at least one The processor is executed, so that at least one processor can perform the method as described above.

According to an embodiment of the present disclosure, there is a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the method as described above.

According to an embodiment of the present disclosure, a computer program product includes a computer program, and the computer program implements the method as described above when executed by a processor.

FIG. 11 shows a schematic block diagram of an example electronic device 1100 that may be used to implement embodiments of the present disclosure. Electronic device is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. Electronic devices may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are by way of example only, and are not intended to limit implementations of the disclosure described and/or claimed herein.

As shown in FIG. 11 , the device 1100 includes a computing unit 1101 that can be executed according to a computer program stored in a read-only memory (ROM) 1102 or loaded from a storage unit 1108 into a random-access memory (RAM) 1103. Various appropriate actions and treatments. In the RAM 1103, various programs and data necessary for the operation of the device 1100 can also be stored. The computing unit 1101 , ROM 1102 , and RAM 1103 are connected to each other through a bus 1104 . An input/output (I/O) interface 1105 is also connected to the bus 1104 .

Multiple components in the device 1100 are connected to the I/O interface 1105, including: an input unit 1106, such as a keyboard, a mouse, etc.; an output unit 1107, such as various types of displays, speakers, etc.; a storage unit 1108, such as a magnetic disk, an optical disk, etc. ; and a communication unit 1109, such as a network card, a modem, a wireless communication transceiver, and the like. The communication unit 1109 allows the device 1100 to exchange information/data with other devices through a computer network such as the Internet and/or various telecommunication networks.

The computing unit 1101 may be various general-purpose and/or special-purpose processing components having processing and computing capabilities. Some examples of computing units 1101 include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), various dedicated artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processing processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1101 executes various methods and processes described above, such as a table generation method. For example, in some embodiments, the table generation method may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 1108 . In some embodiments, part or all of the computer program may be loaded and/or installed on the device 1100 via the ROM 1102 and/or the communication unit 1109 . When the computer program is loaded into the RAM 1103 and executed by the calculation unit 1101, one or more steps of the above-described table generation method may be performed. Alternatively, in other embodiments, the calculation unit 1101 may be configured to execute the table generation method in any other suitable manner (for example, by means of firmware).

Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips Implemented in a system of systems (SOC), complex programmable logic device (CPLD), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include being implemented in one or more computer programs executable and/or interpreted on a programmable system including at least one programmable processor, the programmable processor Can be special-purpose or general-purpose programmable processor, can receive data and instruction from storage system, at least one input device, and at least one output device, and transmit data and instruction to this storage system, this at least one input device, and this at least one output device an output device.

Program codes for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special purpose computer, or other programmable data processing devices, so that the program codes, when executed by the processor or controller, make the functions/functions specified in the flow diagrams and/or block diagrams Action is implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media would include one or more wire-based electrical connections, portable computer discs, hard drives, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, compact disk read only memory (CD-ROM), optical storage, magnetic storage, or any suitable combination of the foregoing.

To provide for interaction with the user, the systems and techniques described herein can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user. ); and a keyboard and pointing device (eg, a mouse or a trackball) through which a user can provide input to the computer. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and can be in any form (including Acoustic input, speech input or, tactile input) to receive input from the user.

The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., as a a user computer having a graphical user interface or web browser through which a user can interact with embodiments of the systems and techniques described herein), or including such backend components, middleware components, Or any combination of front-end components in a computing system. The components of the system can be interconnected by any form or medium of digital data communication, eg, a communication network. Examples of communication networks include: Local Area Network (LAN), Wide Area Network (WAN) and the Internet.

A computer system may include clients and servers. Clients and servers are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, a server of a distributed system, or a server combined with a blockchain.

It should be understood that steps may be reordered, added or deleted using the various forms of flow shown above. For example, each step described in the present disclosure may be executed in parallel, sequentially, or in a different order, as long as the desired result of the technical solution disclosed in the present disclosure can be achieved, no limitation is imposed herein.

The specific implementation manners described above do not limit the protection scope of the present disclosure. It should be apparent to those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made depending on design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present disclosure shall be included within the protection scope of the present disclosure.

Claims (20)

1. A form generation method, comprising:

acquiring text content information and layout information of a target page, wherein the layout information of the target page comprises page parameter information, cell layout information and text layout information;

obtaining cell coordinate information according to the page parameter information, the cell layout information and the text layout information;

obtaining text coordinate information according to the text layout information and the cell coordinate information; and

And generating a target table and labeling data corresponding to the target table according to the cell coordinate information, the text coordinate information and the text content information.

2. The method of claim 1, wherein the cell layout information includes size proportion information of cells and relative positional relationship information between cells; the page parameter information and the cell layout information obtain cell coordinate information, including:

obtaining minimum cell size information according to the page parameter information and the cell size proportion information;

obtaining cell size information according to the minimum cell size information and the relative position relation information; and

and obtaining the cell coordinate information according to the cell size information, the page parameter information and the text layout information.

3. The method of claim 2, wherein the text layout information includes line feed identification information of text within a cell, and the obtaining the cell coordinate information according to the cell size information, the page parameter information, and the text layout information includes:

Identifying the line feed identification information to obtain the line number information of the arrangement of the characters in the unit cells;

obtaining arrangement position information of the cells on the target page according to the cell size information and the page parameter information; and

and obtaining the cell coordinate information according to the arrangement position information and the arrangement line number information.

4. The method of claim 3, wherein the obtaining the cell coordinate information according to the arrangement position information and the arrangement line number information includes:

obtaining a plurality of target cells positioned in the same row according to the arrangement position information;

obtaining the height information of the plurality of target cells according to the arrangement line number information; and

and obtaining the cell coordinate information according to the arrangement position information and the height information.

5. The method of claim 4, wherein the obtaining the height information of the plurality of target cells according to the arrangement line number information includes:

ordering the arrangement line number information of the plurality of target cells to obtain an ordering result; and

and obtaining the height information according to the sequencing result.

6. The method of claim 1, wherein the text layout information includes layout information of text in a cell, text size information, and line feed identification information of text in a cell, and the obtaining text coordinate information according to the text layout information and the cell coordinate information includes:

determining a target strategy according to the arrangement mode information;

obtaining the arrangement line number information of the characters in the unit cells according to the line feed identification information; and

based on the target strategy, the text coordinate information is obtained according to the unit cell coordinate information, the text size information, the arrangement mode information and the arrangement line number information.

7. The method of claim 6, wherein the text size information includes text height information and text width information, the obtaining the text coordinate information based on the target policy from the cell coordinate information, the text size information, the arrangement information, and the arrangement number information includes:

obtaining character abscissa information according to the cell coordinate information, the character width information and the arrangement mode information;

Based on the target strategy, acquiring text ordinate information according to the cell coordinate information, the text height information and the arrangement line number information; and

and obtaining the text coordinate information according to the text abscissa information and the text ordinate information.

8. The method of claim 7, wherein the obtaining text abscissa information according to the cell coordinate information, the text width information, and the arrangement information includes:

according to the arrangement mode information, determining offset parameter information of the characters in the horizontal direction; and

and obtaining the text abscissa information according to the offset parameter information, the text width information and the cell coordinate information.

9. The method of claim 8, wherein the cell coordinate information includes left vertex coordinate information and right vertex coordinate information, the obtaining the text abscissa information based on the offset parameter information, the text width information, and the cell coordinate information includes:

obtaining cell width information according to the left vertex coordinate information and the right vertex coordinate information;

Under the condition that the arrangement mode information is determined to be aligned in the middle, acquiring the character abscissa information according to the left vertex coordinate information, the cell width information and the character width information;

under the condition that the arrangement mode information is determined to be left side alignment, acquiring the text abscissa information according to the left vertex coordinate information and the offset parameter information; and

and under the condition that the arrangement mode information is right side alignment, acquiring the character abscissa information according to the right vertex coordinate information, the character width information and the offset parameter information.

10. The method of claim 7, wherein the cell coordinate information includes upper left vertex coordinate information of a cell, the obtaining text ordinate information based on the target policy according to the cell coordinate information, the text height information, and the arrangement line number information includes:

obtaining minimum cell height information according to the cell coordinate information; and

and obtaining the ordinate information of the characters according to the coordinate information of the upper left vertex, the height information of the minimum cell, the height information of the characters and the information of the number of rows of arrangement.

11. The method of claim 10, wherein the arrangement line number information includes arrangement line number information of characters in a cell and maximum arrangement line number information of characters in a plurality of cells of the same line; the obtaining the text ordinate information according to the upper left vertex coordinate information, the minimum cell height information, the text height information and the arrangement line number information comprises the following steps:

under the condition that the arrangement mode information is determined to be top aligned, acquiring the text ordinate information according to the upper left vertex coordinate information, the minimum cell height information, the text height information and the maximum arrangement line number information; and

and under the condition that the arrangement mode information is determined to be aligned in the middle or aligned at the bottom, acquiring the ordinate information of the characters according to the coordinate information of the top left vertex, the height information of the minimum cell, the height information of the characters, the information of the number of rows of the characters in the cell and the information of the maximum number of rows of the characters.

12. The method of claim 1, wherein the generating a target form and annotation data corresponding to the target form from the cell coordinate information, the text coordinate information, and the text content information comprises:

Determining a target cell according to the text content information;

filling the text content information into a target cell according to the text coordinate information;

arranging the target cells on the target page according to the cell coordinate information to generate the target table; and

and processing the cell coordinate information, the text coordinate information and the text content information according to a preset data format to obtain the labeling data.

13. The method of claim 1, further comprising:

obtaining a linear template of a table frame; and

and constructing frame lines on the target table according to the table frame line type template and the cell coordinate information to obtain a first table image.

14. The method of claim 13, further comprising:

acquiring a seal material image set;

randomly selecting a target seal image from the seal material image set; and

and processing the target seal image and the first form image to obtain a second form image.

15. The method of claim 14, wherein the processing the target stamp image and the first form image to obtain a second form image comprises:

Cutting the target seal image to obtain a local seal image;

deleting the background color in the local seal image to obtain a target local seal image; and

and processing the target local seal image and the first form image to obtain a second form image.

16. The method of claim 15, wherein the processing the target partial stamp image and the first form image to obtain a second form image comprises:

the target local seal image and the first form image are fused, so that a form image with the target local seal is obtained; and

and carrying out filtering treatment on the form image with the target local seal to obtain the second form image.

17. A form generation apparatus comprising:

the first acquisition module is used for acquiring text content information and layout information of a target page, wherein the layout information of the target page comprises page parameter information, cell layout information and text layout information;

the first obtaining module is used for obtaining the cell coordinate information according to the page parameter information, the cell layout information and the text layout information;

The second obtaining module is used for obtaining the text coordinate information according to the text layout information and the cell coordinate information; and

and the generating module is used for generating a target table and marking data corresponding to the target table according to the cell coordinate information, the text coordinate information and the text content information.

18. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-16.

19. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-16.

20. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any of claims 1-16.

Images