JP2001143018A - Character reader and method therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately detect cells of an item column from a table format document, including a header and a footer and to accurately read out characters in cells of a data column corresponding to the detected cells. SOLUTION: A cell coordinate detection part 14 extracts a ruled line from the image data of a table format document stored in an image memory 12 and finds out the coordinate positions of respective cells. A table-aligning part 18 aligns cells included in each row and each column and generates table structure data representing the structure of a table formed by respective cells. A row sorting part 20 detects a section, in which rows having the same number of cells are longest continued in the column direction on the basis of the table structure data, compares the leading row with its upper row by using a difference between the widths of the rows, an interval between the rows, etc., and detects a row to be used as an item row. Then a character recognition part 22 reads out characters from the cells of the item row, recognizes the attributes of characters in the cells of the corresponding data row on the basis of the read characters and converts the attributes of the characters into a suitable character code.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character reading apparatus for reading characters from a tabular document, and more particularly to a suitable character reading apparatus for electronically reading desired characters written on a document such as a form. It relates to a character reading method.

[0002]

2. Description of the Related Art Conventionally, ruled lines are extracted from a table format document,
As a character reading device for reading characters in a cell surrounded by ruled lines, for example, a device described in Japanese Patent Publication No. 2740335 has been proposed.

The character reading apparatus includes an input unit for inputting a tabular document as image data, a storage unit for storing the image data, and extracting a ruled line from the stored image data to recognize a cell surrounded by the ruled line. A cell extracting unit, an item column character recognizing unit that recognizes a character in a cell of a cell corresponding to an item column in the cell, a cell attribute determining criterion storing unit that stores a criterion for determining a cell attribute, It has a cell attribute determining unit that determines the cell attribute of the item column other than the item column from the characters in the item column based on the determination criterion, and a character recognition unit that recognizes characters in each cell according to the determined cell attribute.

According to this apparatus, as described in the above-mentioned patent publication, a tabular document such as a numerical table or a list, which has an item field on the first line of a single table, is used. As an object, characters in a data column following the item column are read based on respective item characters in the item column.

That is, the character reading device scans a tabular document at an input section and reads image data representing the resulting image. The loaded image data is
The data is sequentially stored in the storage unit. When the image data is accumulated, the cell extracting unit extracts the horizontal ruled line and the vertical ruled line, and obtains the coordinate position of the cell surrounded by the ruled line. When each cell is extracted by the cell extraction unit, the image data of the character in the cell is read out from the storage unit from the coordinate position of each cell in the first row by the item column character recognition unit, and each character is read. recognize.

Next, upon recognizing the character of each cell in the item column, the cell attribute determining unit determines the cell attribute of the character in the item column based on the criterion of the cell attribute criterion unit. Determine the cell attribute of the data column. The result of the determination is set in the character recognition unit, and the characters in the corresponding cell are read in the data column following the item column, and each character is converted into a character code.

[0007]

However, in the above-described conventional technique, the characters written in the cells in the first row of the table extracted by the cell extracting unit are used by the item column recognizing unit to respectively write the characters in the item columns. If the table format document contains a header or footer surrounded by ruled lines other than the target table, the item column cannot be recognized correctly, and the character recognition in the data column will not be recognized. There was a problem that it could come.

When the item column is represented by a plurality of rows of cells different from the data column, for example, when the item column is partitioned over two rows of cells, a correct item for reading the data column is used. In some cases, the field cannot be recognized, and in the same manner as described above, there is a problem that the character recognition in the data field may be affected.

Therefore, when a tabular document has a header or footer, or when an item column extends over a plurality of lines, a reading area must be specified by an operator or the like, which is troublesome.

An object of the present invention is to solve the above-mentioned problems and to provide a character reading device capable of accurately detecting an item column without human intervention and reading characters in a required data column accurately. With the goal.

[0011]

According to the present invention, there is provided a character reading apparatus for reading a character written in a predetermined cell surrounded by a ruled line from a tabular document. Image input means for inputting image data representing an image of a formal document, image data storage means for storing image data from the image input means, and image data of ruled lines extracted from the image data in the image data storage means. Cell coordinate detecting means for obtaining ruled area data representing coordinate information of each cell surrounded by ruled lines, and aligning the cells in the row direction and column direction based on the ruled area data obtained by the cell coordinate detecting means. Means for generating table structure data representing a table structure formed by cells of each row and column; and table structure data generated by the table alignment means. A row classifying means for classifying each row into a cell of an item row, a cell of a data row, and other cells based on the data, and characters written in at least each cell of the data row classified by the row classifying means. A character recognition unit that reads from the data storage unit, converts the character code into a character code, and recognizes a character of a cell in a data column corresponding to each item based on a character attribute in a cell written in the item row. And

In this case, the row classifying means includes a cell number detecting means for calculating the number of cells included in each row based on the table structure data generated by the table sorting means, and a row having the same cell number being the longest in the column direction. A continuous section detecting means for detecting a continuous section, and an item row detection for detecting a true item row from a result of comparing the temporary item row with a row further above, with the first row of the section as a temporary item row Means.

In the item line detecting means, the provisional item line and the line above it are compared, and if the difference in line width, the line interval, or both are within predetermined values,
The upper row is set as a new provisional item row and compared with the further upper row, and if each exceeds a predetermined value, it is advantageous that the provisional item row is a true item row.

The item line detecting means compares the provisional item line with the line above it, and determines the difference in the cell area included in the line or the difference in the ratio of the cell area to the line area, and If the interval is within a predetermined value, the upper row is regarded as a new temporary item row and compared with the row above it, and if each exceeds the predetermined value, the temporary item row is true. The item line may be detected as the item line.

In these cases, the table sorting means determines to which row the cell in the row direction belongs, based on the coordinate position and the height of each cell, and sorts the cells in each row. Good.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, an embodiment of the character reading device according to the present invention will be described in detail with reference to the accompanying drawings. FIG.
1 shows an embodiment of the character reading device according to the present invention. The character reading device according to the present embodiment is a character reading device that extracts a ruled line from a tabular document such as a form K shown in FIG. 3 and reads characters in a predetermined cell surrounded by the ruled line.
From the tabular document including the header B and the footer C surrounded by ruled lines in addition to the table body A to be read, each character written in the cell of the data column of the table body A is read, and a predetermined character code is set. This is a data conversion device that converts the data into.

In particular, in the present embodiment, cells surrounded by ruled lines are aligned in the row and column directions to generate table structure data representing the structure of a table formed by each cell.
1 includes a table sorting unit 18 shown in FIG. 1 and a row classifying unit 20 for classifying item rows and data rows from the table body A based on the result.
The main feature is that the characters in the corresponding data cells are effectively converted into character codes based on the characters written in the cells of the classified item rows.

More specifically, as shown in FIG. 1, the character reading apparatus according to the present embodiment includes an image input unit 10, an image memory 12, a cell coordinate detecting unit 14, a table cell information memory 16, Part 18
, A line classification unit 20, a character recognition unit 22, and a recognition result memory 24.
, A data processing unit 26, a display unit 28, an operation input unit 30,
And a control unit 32.

The image input unit 10 is an input device for inputting image data representing an image image on a tabular document. In this embodiment, the image input unit 10 scans a target tabular document and converts its image signal into a binary signal. An image scanner that reads as image data is advantageously applied. The input image data is sequentially written to the image data memory 12. The image data memory 12 is a storage circuit that sequentially stores the image data from the image input unit 10, and is a storage device such as a frame memory that expands the input image data into a bitmap and stores it. The stored image data is supplied to the cell coordinate detector 14.

The cell coordinate detecting section 14 includes an image data storing section 12
Is a cell extraction circuit that detects the coordinate position of each cell surrounded by ruled lines based on the image data accumulated in
For example, the image data "1" detects a solid line and a ruled line represented by an intermittently continuous broken line continuous for a predetermined length or more in the row direction or the column direction, and calculates the coordinates of each cell from the intersection point. This is a coordinate detection circuit to be obtained.
Information indicating the coordinate position of each cell is sequentially written to the table cell information memory 16 as ruled line area data and supplied to the table aligning unit 18.

The table aligning unit 18 is a table structure detecting circuit for obtaining a table structure formed by each cell based on the ruled line area data obtained by the cell coordinate detecting unit 14. Is determined to be included in which row and column, and the cells are aligned in the row and column directions to generate table structure data representing the table structure.
In particular, in this embodiment, when cells are aligned in the row direction,
Based on the coordinate position of each cell and its height,
Decide which row it belongs to and align the cells in each row. The generated table structure data is sequentially written to the table cell information memory 16 and supplied to the row classification unit 20.

The row classifying unit 20 is a cell classifying circuit for classifying cells of each row into cells of an item row, cells of a data row and other cells based on the table structure data generated by the table sorting unit 18. In the present embodiment, in this embodiment, the section where the row having the same number of cells is the longest in the column direction is detected, and the first row of the section and the row further above are compared to obtain an item row. Subsequent to this, the data is classified into data rows of the same cell number section and other header and footer rows.

More specifically, the line classification unit 20 of the present embodiment
Is a cell number detecting section for respectively obtaining the number of cells included in each row from the table structure data, a continuous section detecting section for detecting a section where the row having the same cell number is the longest continuous in the column direction from the result, And an item line detection unit that detects a true item line from the result of comparing the tentative item line with a line above it as a tentative item line. In particular, when comparing the tentative item line and the line above it, the item line detection unit of the present embodiment determines whether the difference between the widths of those lines is within a predetermined value, and furthermore, Is determined whether or not the interval is within a predetermined value, and when each is within the predetermined value, the upper row is a temporary item row, and further compared with the row above,
When each of the values exceeds a predetermined value, a provisional item line is detected as a true item line and the item line is detected. When an item row is detected, each row of the section where the same number of cells is continuous below is a data row, a row above the item row is a header row, and a row below the data row section is a footer row. The result of each classification is written in the table cell information memory 16. The result of the row classification written in the table cell information memory 16 is supplied to the character recognition unit 22 together with the ruled line area data and the table structure data.

A character recognizing section 22 reads out from the image memory 12 the image data of the characters written in each cell of the data row classified by the row classifying section 20, and converts the image data into a predetermined character code. In the present embodiment, the characters in the cells of the corresponding data row are recognized based on the character attributes in the cells of the item row classified by the row classification unit 20. More specifically, an item line character recognition unit that reads image data of characters in each cell of an item line from the image memory 12 and converts it into a character code, and a cell that stores a criterion for determining a cell attribute An attribute determination criterion storage unit, a cell attribute determination unit that determines a cell attribute of a data row from a character of an item row based on the determination criterion, and image data representing a character in a cell of each data row according to the determined cell attribute. And a data line character recognizing unit which reads out from the memory 12 and converts it into a character code. The converted character codes are respectively written in the recognition result memory 24.

The recognition result memory 24 is a storage circuit for accumulating the character of each cell converted into the character code for each data line. The result is converted into the character corresponding to the character code and displayed on the display unit 28. Is done. Advantageously, the data is displayed after being subjected to processing such as rearrangement or numerical calculation in the data processing unit 26. The display unit 28 displays a character represented by the character code from the recognition result memory 24 (CRT (cathode ray)).
The present embodiment displays a recognition result of character reading and a result of data processing thereof. The operation such as the display switching is executed in accordance with an instruction from the operation input unit 30.

The operation input unit 30 is an operation circuit which is connected to a keyboard or a pointing device, and is used for switching display or instructing data processing or inputting desired data. The operation instruction is supplied to the control unit 32. The control unit 32 is a central processing unit that controls the sequence control of character reading, data processing, and the like. In particular, in this embodiment, in response to image data input of the image input unit 10, the cell coordinate detection unit 14 and the table This is a sequence control circuit that sequentially controls the sorting unit 18, the line classification unit 20, and the character recognition unit 22.

Next, the operation of the character reading apparatus according to the present embodiment will be described with reference to FIGS. First, as shown in the flowchart of FIG. 2, in step S10, the image input unit 10 scans a tabular document, reads the image as binary image data, and sequentially stores the binary image data in the image memory 12.
For example, in the present embodiment, as shown in FIG. 3, as a tabular document, a document image of a form K including a header B and a footer C surrounded by ruled lines in addition to a table body A in which data to be read is written. Is read.

Next, when the image data is stored in the image memory 12, the process proceeds to step S12, in which the cell coordinate detection unit 14 reads out the image data from the image memory 12, extracts the image data of each ruled line, The coordinate position of each cell surrounded by the ruled line is obtained. As a result, the obtained cell coordinate information is sequentially written into the table cell information memory 16 as ruled line area data.

Next, when the coordinate position of the cell is obtained, the process proceeds to step S14, in which the table aligning unit 18 reads out the ruled line area data from the table cell information memory 16 and, based on the coordinate information, writes the cells of each row in the row direction. Align and find which cells are in which row. At this time, by arranging the respective rows sequentially from the first row based on the position coordinates of the cells in the row direction, for example, as shown in FIG. 4, in the third row, six cells 100 to 110 are aligned. Thus, the cell 112 corresponding to the cell below the second column is detected as the fourth row. Therefore, in the present embodiment, the cells are aligned in the row direction based on the height together with the position coordinates of the cells. As a result, as shown in FIG. 5, the cell 112 is obtained as the seventh cell included in the third row.

Similarly, the cells in each row are aligned in the row direction based on the position coordinates and the height. When the alignment in the row direction is completed, the cells are aligned in the column direction in step S16 by calculating the spacing between the rows. The table structure data representing the table structure formed by the cells of each row and column thus obtained is written to the table cell information memory 16.

Next, the process proceeds to step S18, where the line classification unit 20
The table structure data is read from the table cell information memory 16 to determine the number of cells included in each row as shown in FIG. After calculating the number of cells in each row, step S2
Proceeding to step 0, the section D having the cell number "10" from the fourth row to the eleventh row as shown in FIG. 6 is detected from the result of step S18, for example, as shown in FIG. .

Next, the process proceeds to step S22, and the process proceeds to step S20.
The head row of the section D detected in the above is set as a temporary item row E as shown in FIG. 7, and the process proceeds to step S24. In step S24, as shown in FIG. 8, the difference between the width of the provisional item line E and the width of the line F above it is determined, and whether or not the value is within a predetermined value is determined. Is determined to be within a predetermined value. If the results are within the predetermined values, the process proceeds to step S26, where the upper row F is displayed as shown in FIG.
Is returned to step S24 as a tentative item row, and the upper row F, which is the tentative item row, is further compared with the upper row G. As a result, if each exceeds a predetermined value, step
Proceeding to S28, the item line F is detected as shown in FIG. 10 with the provisional item line F as a true item line.

Next, when the item line F is detected, the step
At S30, each row in section D where the same number of cells below the item row is continuous is used as a data row, and each row above the item row is used as a header row and each row further below data row section D is used as a footer row. Classify as The results of each classification are written to the table cell information memory 16. The result of the line classification written in the table cell information memory 16 is supplied to the character recognition unit 22 together with the ruled line area data and the table structure data.

Next, when the line classification is completed, step S32
Then, from the coordinate position of each cell in the row classified as the item row by the character recognition unit 22, image data corresponding to characters in those cells is read from the image memory 12 to recognize each character. Next, in step S34,
The cell attribute of the data row is determined from the characters of the item row based on the criteria stored in advance. As a result, the step
In S36, the image data representing the character in the cell of each data row corresponding to the cell of the item row is read from image memory 12, and each character is converted into a character code according to the cell attribute determined in step S34. The converted character codes are respectively written in the recognition result memory 24.

Next, in step S38, the characters of each data line written in the recognition result memory 24 are displayed on the display unit 28, and it is confirmed whether or not each character has been correctly read. If each character is valid, the process proceeds to step S40, where the data processing unit 26 is started in response to the operation of the operation input unit 30, and the desired data processing such as rearrangement or numerical calculation is performed. Ends.

Similarly, image data representing the document image of each tabular document is read, and ruled lines are extracted from the image data, and the coordinate position of a cell surrounded by each ruled line is obtained. Then, aligning the cells in the row and column directions based on the coordinate position and height of each cell,
Based on the result, cells in the item row, cells in the data row, and other cells are classified. In particular, when detecting an item row, a section where the same number of cells is the longest and continues in the column direction is obtained, and the item row is detected from the result of comparing the top row with the upper row and the upper row. When an item row is detected, the character in the cell is recognized, the character in the data row is recognized based on the result, and each is converted to a character code. As a result, the characters written in the cells of the respective required data rows written in the table body A of the tabular document are accurately read, and the desired data processing is executed.

As described above, according to the character reading apparatus of the present embodiment, each cell is aligned in the row direction and the column direction by the table aligning section 18 to represent the table structure formed by each cell. Based on the table structure data, based on the table structure data, the row classification unit 20 detects a section where rows of the same number of cells are continuous in the column direction, and determines the first row, the upper row, and the upper row. Since the item line is detected by comparison, even if the tabular document has a header or footer surrounded by ruled lines, the item line can be accurately detected, and as a result, based on the character of the item line, Characters written in the data line can be accurately converted to character codes and read.

When the cells in the row direction are sorted by the table sorting unit 18, it is determined which cell the cell belongs to based on the position coordinates and the height of each cell in the row direction. Since the rows are arranged, even if the item row is divided into cells of multiple columns, each cell of the item row can be detected accurately, and as a result, data can be detected based on the character of the item row. Characters written on a line can be accurately converted to character codes and read.

Therefore, even when a tabular document has a header or footer, or when an item row extends over a plurality of rows, the characters in the required data fields can be accurately specified without designating a reading area by an operator or the like. Can be read.

In the above embodiment, when comparing the provisional item line with the line above it, the line classifying section 20 detects the item line based on the difference between the line widths and the line interval. However, in the present invention, when comparing the provisional item row with the row above it, the area ratio of the cell area to the cell area or the row area contained in those rows and the row interval are compared. Then, the item row may be detected based on whether or not the difference is within a predetermined value.

In the above embodiment, the case where an image scanner for scanning a tabular document and forming its binary image data is applied as the image input unit 10 has been described as an example. The circuit is not limited to this, and may be a circuit that receives and inputs an image signal from a transmission device such as a facsimile. Furthermore, in the present invention, a circuit for inputting image data including gradation or color image data, for example, is not limited to binary image data, as long as the image data can accurately read ruled lines and characters. It may be.

Further, in the above embodiment, for the sake of simplicity, the case where the table format documents are read one by one and the respective characters are read has been described as an example. You may read and process each.

[0043]

As described above, according to the present invention, each cell is aligned in the row direction and the column direction based on the ruled line area data representing the coordinate information of the cell surrounded by the ruled line, and each row is aligned. Table aligning means for generating table structure data representing a table structure formed by and table cells, and a cell of an item row and a cell of a data row based on the table structure data generated by the table aligning means and other cells. Since the cell has a row classification means for classifying each row, even if there is a header or footer surrounded by ruled lines other than the target table in the tabular document, the item row can be accurately detected. The characters in the data line can be accurately converted to character codes based on the characters in the item line. Therefore, it is possible to omit trouble such as designating a reading area by an operator or the like, and it is possible to obtain an excellent effect that processing can be efficiently executed in a short time.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of a character reading device according to the present invention.

FIG. 2 is a flowchart for explaining the operation of the character reading device according to the embodiment of FIG. 1;

FIG. 3 is a diagram showing an example of a tabular document read by the character reading device according to the embodiment of FIG. 1;

4 is a table aligning unit 18 of the character reading device according to the embodiment of FIG.
It is a figure which shows the process of processing.

FIG. 5 is a table aligning unit 18 of the character reading device according to the embodiment of FIG. 1;
It is a figure showing the processing result of.

FIG. 6 is a line classification unit 20 of the character reading device according to the embodiment of FIG. 1;
FIG. 6 is a diagram showing detection of a continuous section by using FIG.

FIG. 7 is a line classification unit 20 of the character reading device according to the embodiment of FIG. 1;
FIG. 6 is a diagram showing a determination of a provisional item line according to FIG.

FIG. 8 is a line classification unit 20 of the character reading device according to the embodiment of FIG. 1;
FIG. 7 is a diagram showing an example of a comparison process of a provisional item row according to FIG.

FIG. 9 is a line classification unit 20 of the character reading device according to the embodiment of FIG. 1;
FIG. 7 is a diagram showing an example of a comparison process of a provisional item row according to FIG.

FIG. 10 is a line classification unit of the character reading device according to the embodiment of FIG. 1;
FIG. 7 is a diagram showing a true item row determined by 20.

[Explanation of symbols]

 10 Image input unit 12 Image memory 14 Cell coordinate detection unit 16 Table cell information memory 18 Table alignment unit 20 Line classification unit 22 Character recognition unit

Claims (10)

[Claims] 1. A character reading apparatus for reading a character written in a predetermined cell surrounded by a ruled line from a tabular document,
The apparatus includes: image input means for inputting image data representing an image of a tabular document; image data storage means for storing image data from the image input means; and a ruled line based on the image data stored in the image data storage means. Cell-coordinate detecting means for extracting image data of (a) and calculating ruled-line area data representing coordinate information of each cell surrounded by the ruled line; and determining each cell based on the ruled-line area data obtained by the cell coordinate detecting means. Are aligned in the row and column directions,
Table sorting means for generating table structure data representing a table structure formed by cells of each row and column; cells of item rows and cells of data rows based on the table structure data generated by the table sorting means And a row classifying means for classifying each row into cells other than those, reading characters written in each cell of at least the data row classified by the row classifying means from the image data storage means into a character code. A character reading device, comprising: character recognition means for recognizing a character in a cell of a data line corresponding to each item based on an attribute of a character in a cell written in the item line. 2. The character reading device according to claim 1, wherein the row classification unit determines the number of cells included in each row based on the table structure data generated by the table alignment unit; A continuous section detecting means for detecting a section in which a row having the same number of cells is the longest and continuous in the column direction, and a result of comparing the provisional item row with a further upper row using the first row of the section as a provisional item row And an item line detecting means for detecting a true item line from the character reading device. 3. The character reading device according to claim 2, wherein the item line detection unit compares the provisional item line with a line above the tentative item line, and determines a difference in line width, a line interval, or any of those. If both are within the respective predetermined values, the above line is regarded as a new provisional item line and compared with the line above it, and if each exceeds the predetermined value, the provisional item line is A character reading device characterized by being a true item line. 4. The character reading device according to claim 2, wherein the item line detection unit compares the provisional item line with a line above the tentative item line, and determines a difference in the area of cells included in the line or a difference in the line. If the difference between the ratio of the cell area to the area and the row spacing is within a predetermined value, the above row is set as a new provisional item row and compared with the row above it, and The character reading device according to claim 1, wherein when the value exceeds a predetermined value, the provisional item line is regarded as a true item line. 5. The character reading device according to claim 1, wherein in the table aligning unit, the cell in the row direction is determined based on a coordinate position of each cell and its height. A character reading device, wherein it is determined whether or not a cell belongs to a row, and cells in each row are arranged. 6. A character reading method for reading characters written in predetermined cells surrounded by ruled lines from a tabular document,
The method comprises the steps of: inputting image data representing an image of a tabular document by image input means; storing image data from the image input means in image data storage means; and storing the image data in the image data storage means. An extracting step of extracting image data of a ruled line from the stored image data to obtain ruled line area data representing coordinate information of each cell surrounded by the ruled line; An alignment step of aligning in the column direction to generate table structure data representing a table structure formed by cells of each row and column; and a cell of an item row and a cell of a data row based on the table structure data; A classification step of classifying each row into cells other than the above, and at least the characters written in each cell of the data row among the classified cells are described above. Recognizing a character in a cell in a data line corresponding to each item based on an attribute of a character in a cell written in the item line by reading from the image data storage unit and converting the character code into a character code. A character reading method characterized by the following. 7. The character reading method according to claim 6, wherein the classifying step further obtains the number of cells included in each row based on the table structure data, and the row having the same number of cells is the longest in the column direction. A data line detecting step of detecting a section that is continuous with the above, and an item row for detecting a true item row from a result of comparing the tentative item row and a row further above with the first row of the section as a tentative item row And a detecting step. 8. The character reading method according to claim 7, wherein the item line detecting step further compares the tentative item line and a line above the tentative item line to determine a difference in a line width, a line interval, or the like. If both are within a predetermined value, the above line is further compared with a line above it as a new provisional item line, and if each exceeds a predetermined value, the provisional item line is true. A character reading method including a comparison step of setting an item row. 9. The character reading method according to claim 7, wherein the item line detecting step further compares the provisional item line with a line above the tentative item line to determine a difference in cell area or a line included in the line. If the difference between the ratio of the cell area to the area of the cell and the row interval is within a predetermined value, the above row is compared with a row above it as a new provisional item row, and each , A character reading method including a comparison step of setting a provisional item line as a true item line. 10. The character reading method according to claim 6, wherein the aligning step of generating the table structure data further includes a coordinate position of each cell in a row direction and a height thereof. A character reading method comprising: determining a row to which a cell belongs based on the position, thereby aligning cells in each row.