JP4503520B2 - Document data reading apparatus and computer program therefor - Google Patents

Description

  The present invention relates to a document data reading device that reads out input document data with synthesized speech and a computer program thereof.

  A tabular format is often used as one of the methods of expressing information that is simply transmitted to the user. The tabular format is very effective in that it can visually indicate information, that is, the association and conversion of elements specified from the items, by expressing them in two dimensions, vertical items and horizontal items. Is the transmission format. In addition, a tabular format is often used in websites that disclose various information via the Internet.

  By the way, as a conventional technique for assisting in transmitting information displayed on the website as described above to a visually handicapped person, there are a software program for reading a voice, a screen reader, and the like. However, the speech-to-speech software program searches for characters expressed in a table format from the upper left to the lower right of the table, and reads out the characters as a one-dimensional sequence in the order they appear. In other words, after reading the first column from left to right, the second column is read from left to right and then the third column, so information as an element corresponding to two-dimensional items of vertical and horizontal Rather, it is simply a list of one-dimensional information, and even if a visually impaired person hears a voice that is read out, it will be read out at the beginning of reading out what is the row item and column item of the information currently being read out. It ’s important to go back and understand what the items in the row and column are, what the relationship is between the items in the row and the column, and what information each element corresponds to which item in the row and the column. It is very difficult. In addition, for large tables composed of many rows and columns, or tables that do not have one-to-one correspondence between row and column items, what items exist in the rows and columns? It becomes more difficult to grasp.

  Also, instead of the above table structure itself, for example, visual information such as white and black circles representing wins and losses, and “numbers-numbers” representing the score of the competition results should be read aloud by a speech-to-speech software program. It is difficult to understand by listening.

  In Patent Document 1, focusing on the description of data included in a table representing scores and the like, focusing on the baseball win / loss table and the sumo wrestler table, “(numerical string A) + (horizontal bar) + (numerical string B ) ”As“ A vs. B ”, and the natural language comes before the number A and after the number B, the natural language is determined as the team name, and the natural language corresponding to the team with a large number is We are proposing a technology that reads out with "win". In addition, when supplementary information such as game location and game start time is added with parentheses, we propose a method to convert the table information to natural language by setting rules in the dictionary, such as not reading it as unnecessary information. ing.

  However, the technique described in Patent Document 1 focuses on the expression format of information included in each cell of the table, and the reading order is one column from the left to the right as in the conventional voice reading software program. It is a structure that reads out in the order of the first, second, and so on, what the items of the row and column are, the item in which two cells are integrated into one, or conversely, one cell is decomposed into two There is a problem that even if there are other items, the relationship between the items cannot be expressed by reading.

  In the techniques described in Patent Document 2 and Patent Document 3, in order to solve the problem related to the technique described in Patent Document 1, the row and column items corresponding to the information included in each cell of the table are clearly identified. We are trying to make it. That is, the user can select whether to read the cells one row at a time from the right to the left, or to read the cells one column at a time from the top to the bottom by operating the keyboard. It also proposes a function that allows the user to confirm what each item of the row and column is by operating the keyboard when reading of each cell is completed.

  Further, in the techniques described in Patent Document 2 and Patent Document 3, when a plurality of cells are unified into one, the cells are decomposed into minimum units, and the decomposed cells include the same document data and values. It has a function of reading out as a table in which the items of the row and the column correspond one-to-one. It also has a function of reading out only the designated row or column in order. When only a row is read out, it also has a function of reading out what the column items for each cell are.

  Furthermore, in the techniques described in Patent Document 2 and Patent Document 3, it is possible to specify a keyword, extract only the cell including the keyword, and read it out together with the row and column information. When cells having the same content continue, it is one of the features that reading is performed so as not to overlap as in “the same sentence”. Also, it is possible to specify how many rows and columns are read out without reading out the entire table.

As described above, in the techniques described in Patent Document 2 and Patent Document 3, when the necessary information is clear for a user who can understand the information described in the entire table, the table is read out. It is possible to listen to only the necessary information without listening to everything. In addition, the table can be read out by the user's operation, and the row and column information can be confirmed on the spot for each cell, so the user can easily identify the row and column items for each piece of information. It becomes possible to grasp.
JP-A-11-134166 JP 2004-316972 A JP 2000-089778 A

  However, in the techniques described in Patent Documents 2 and 3, it is necessary for the user to know in advance what information is included in the entire table before designating the necessary information. There is a problem that it is difficult for a visually handicapped person who cannot confirm whether or not the information he / she needs exists in the table before listening to the reading.

  SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and its purpose is to provide a document that can transmit information indicating the overall structure of a tabular format when the information expressed in the tabular format is indicated by audio output. To provide a data reading device and a computer program thereof.

In order to solve the above problems, the present invention provides a text data reading device that reads out and outputs input text data by speech synthesis, and the position of a cell in a table composed of one or more rows and one or more columns And input means for inputting sentence data including table information indicating an association with an item included in the cell or an element of the item included in the cell, and the table from the sentence data input by the input means Table detecting means for detecting information, determining means for determining whether each cell of the table information detected by the table detecting means is a cell including the item or a cell including the element, and the determining means Item number calculating means for calculating the number of cells determined as cells including items, element number calculating means for calculating the number of cells determined by the determining means as cells including elements, and item counting A total output means for reading out and outputting by speech synthesis the number of cells calculated by the output means as the number of items, and outputting by reading out the number of cells calculated by the number of elements as a number of elements; and Item output means for reading out and outputting text data included in all cells determined by the determination means as cells including items by speech synthesis; and text data included in the cells determined by the determination means as cells including elements A text data reading device comprising: element output means for reading out by synthesis and outputting.

  The present invention determines whether or not the table information detected by the table detection unit includes an integrated cell in the above-described invention, and determines that the integrated cell is included. The integrated determination dividing means for dividing the divided cells and outputting the table information in which the contents of the items or elements included in the original cell are the contents of all the divided cells. When it is determined that the dividing means does not include an integrated cell, it is determined whether each cell of the table information detected by the table detecting means is a cell including the item or the element, When it is determined that the integrated determination dividing unit includes an integrated cell, it is determined whether each cell of the table information output from the integrated determination dividing unit is a cell including the item or a cell including the element. It is characterized by

  The present invention, in the invention described above, comprises request input means for inputting position information for specifying the position of a cell including the element among the tabular cells in response to a user operation, and the element output The means is characterized in that the text data included in the cell corresponding to the position information input by the request input means is read out by speech synthesis and output.

According to the present invention, in the invention described above, the determination unit is configured such that the table information is a table having items in rows and columns , a table having items only in rows , or a table having items only in columns. If it is determined that the table has items in rows and columns, whether each cell of the table information is a cell including the item corresponding to a row or a cell including the item corresponding to a column If it is determined that the cell includes the element and the table includes an item only in a row, each cell of the table information is a cell including the item corresponding to the row, or a cell including the element determines whether it is, if it is determined that a table with fields only in the column, or each cell of the table information is a cell that contains the item corresponding to the column, determines whether the cell containing the element It is characterized by that.

The present invention relates to a computer of a text data reading device that reads out text data to be input by speech synthesis and outputs the data , and includes a cell position in a table composed of one or more rows and one or more columns, and items included in the cells. Alternatively, input means for inputting text data including table information indicating correspondence with elements of the items included in the cell, and table detection means for detecting the table information from text data input by the input means; Determination means for determining whether each cell of the table information detected by the table detection means is a cell including the item or a cell including the element; and a cell determined by the determination means as a cell including the item An item number calculating means for calculating the number of cells, an element number calculating means for calculating the number of cells determined by the determining means as a cell containing elements, and a cell number calculated by the item number calculating means. The number of cells is read out by voice synthesis as the number of items and output, and the total number of cells calculated by the number-of-elements calculation unit is read out by voice synthesis as the number of elements and output, and the determination unit Item output means for reading out and outputting sentence data included in all cells determined as including cells by speech synthesis, and reading out and outputting sentence data included in the cells determined by the determination means as cells including elements And a computer program for functioning as an element output means.

According to this invention, the document data reading apparatus associates the position of a cell in a table composed of one or more rows and one or more columns with an item included in the cell or an element of an item included in the cell. Text data including table information indicating is input. Table information is detected from the text data to be input, and it is determined whether each cell of the detected table information is a cell including an item or a cell including an element. Calculate the number of cells determined as cells containing an item, calculate the number of cells determined as cells containing an element, read the number of cells calculated as cells containing an item as the number of items, and output it by speech synthesis. The number of cells calculated as a cell including an element is read out by speech synthesis as the number of elements and output. In addition, text data included in all cells determined as cells including items is read out by voice synthesis and output, and text data included in cells determined as cells including elements is read out and output by voice synthesis. . As a result, first, the overall structure of the tabular document data is read out, and further, the contents of each item are read out. The summary of the item can be understood, and then the cell containing the element can be designated and read out.

 Further, according to the present invention, the document data reading apparatus determines whether or not the table information to be detected includes an integrated cell, and determines that the integrated cell is included. Is output, and table information in which the contents of items or elements included in the original cell are the contents of all the divided cells is output. Also, if it is determined that the integrated cell is not included, it is determined whether each cell of the table information to be detected is a cell including an item or a cell including an element, and it is determined that the cell includes an integrated cell. In this case, it is determined whether each cell of the output table information is a cell including an item or a cell including an element. As a result, even in the case of tabular document data in which multiple rows or multiple columns are integrated, by dividing the integrated cells, even in situations where visual usage is difficult, a unified expression can be obtained. It becomes possible to represent the entire image of the tabular document data, and it is possible to easily understand the tabular format in a situation where visual use is difficult.

  Further, according to the present invention, the document data reading apparatus receives position information, inputs position information for designating the position of a cell including an element among tabular cells, and corresponds to the input position information. The text data included in the cell is read out by speech synthesis and output. As a result, after the user understands the overall image of the tabular format, the user can individually specify the cell including the element by the position information and read it out.

Further, according to the present invention, the document data reading apparatus determines whether the table information is a table having items in rows and columns , a table having items only in rows, or a table having items only in columns . If the table is determined to be a table with items in rows and columns, each cell of the table information is a cell that contains items corresponding to rows, cells that contain items corresponding to columns, or cells that contain elements. determine whether, when it is determined that a table with fields only in the row, or a cell that contains the item that each cell of the table information corresponding to the row, determining whether a cell containing the element, the column only When it is determined that the item is a table , it is determined whether each cell of the table information is a cell including an item corresponding to a column or a cell including an element. This makes it possible to read out the overall tabular structure of tabular document data, even for tabular document data other than rows and columns, that is, only rows or columns only, and for each item. Can be read aloud, and in a situation where visual use is difficult, it is possible to easily understand the table format.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a schematic block diagram showing an internal configuration of a document data reading apparatus 1 according to the present embodiment. The input unit 10 inputs document data including a table format including item information and information of one or more elements corresponding to the item. Here, in the following description, each piece of tabular information is described as a cell, a cell including item information is described as a cell including an item, and a cell including element information is described as a cell including an element. The table detection unit 12 detects document data in a tabular format from the document data input from the input unit 10. Here, the detection of the tabular format is, for example, document data published on a website, and in the case of document data described in HTML (Hyper Text Markup Language), the tabular format portion is detected from an HTML tag or the like. To detect. In addition, in the case of document data created with general word processing software, the shape of the ruled line that exists in the sentence (whether there is a horizontal ruled line at the top of the table, there is a delimited ruled line in the item column, The table is detected from the document by a known technique such as analyzing whether the lowest horizontal ruled line exists or the corner ruled line exists.

  The integration determination / division unit 14 determines whether or not a cell in which a plurality of cells are integrated exists in the table format portion detected by the table detection unit 12 based on the overall configuration of the table format. Further, when it is detected by the integration determining / dividing unit 14 that there is a cell in which a plurality of cells are integrated, the cell is divided, and each divided cell generates tabular document data having the contents of the division source cell. Constitute. In the tabular document data, if the upper left cell is composed of blank, diagonal lines, or diagonal lines and rows and columns, the upper left cell is not divided.

  Further, when performing the division, the integration determining / dividing unit 14 records division correspondence information indicating the correspondence between the division source cell and the divided cell in the integrated information storage unit 30.

  The analysis unit 18 includes a determination unit 18a, an item number calculation unit 18b, and an element number calculation unit 18c. When the integrated determination / division unit 14 determines that there is no integrated cell, the integrated determination / If the integrated determination / division unit 14 determines that there is an integrated cell based on the document data in the tabular format input from the division unit 14, the table format after the integration determination / division unit 14 has divided Perform tabular analysis based on document data.

  In the analysis unit 18, the determination unit 18 a determines whether the upper left cell of the input tabular document data is composed of blank, diagonal line, or diagonal line and row and column items. If the cell is composed of blank, slanted line, or slanted line and row and column items, it is determined that the tabular document data is tabular document data composed of rows and columns. In this case, the document data is determined to be tabular document data composed of only rows or columns.

  Further, when the determination unit 18a determines that the document data is in a tabular format including only rows or columns, the determination unit 18a determines how far the cell including the item is, and from where the cell includes the element. Identify the containing cell and the cell containing the element. The determination unit 18a records information specifying the cell including the item and the cell including the element in the data storage unit 32 together with the input document data in the table format.

  The item number calculation means 18b counts and outputs the number of cells determined by the determination means 18a as a cell including an item. At this time, in the case of a table composed of rows and columns, the upper left cell includes items of blanks, diagonal lines, or diagonal lines and rows and columns, and therefore the upper left cells are excluded.

  The element number calculation means 18c counts and outputs the number of cells determined by the determination means 18a as a cell including an element. The count may be calculated by multiplying the number of rows and columns when the document data is tabular document data composed of rows and columns. At this time, the number of rows and columns is the number of rows and columns for a cell including an element, and is not included for a cell including an item.

  The entire output unit 24 receives information specifying the cell including the item and the cell including the element together with the document data in the table format from the determination unit 18a. The overall output unit 24 also includes information specifying the cell including the item and the cell including the element, the number of cells including the item output by the item number calculating unit 18b, and the cell including the element output by the element number calculating unit 18c. Based on the number, the reading data indicating the entire image of the tabular document data is generated, and the generated reading data is reproduced by voice synthesis, that is, read and output. For example, “From the top row is a row (or horizontal) item and there are Δ items. From the left, the □ column is a column (or vertical) item, and there are ◇ items. The number of data of “◎” is “◎”, etc., and the read-out data is reproduced by speech synthesis. In addition, when cells including items exist in a plurality of rows or columns, the number of each item is “○ items in the first row (column), Δ items in the second row, etc.”. Read-out data indicating each item is generated.

  The item output unit 26 receives information specifying the cell including the item and the cell including the element together with the document data in the table format from the determination unit 18a, and reads out the contents of the cell including the item sequentially from 1 Data is generated, and the generated read-out data is reproduced by speech synthesis. If there are rows and columns, “○ items in the row (or horizontal) are 1XX, 2xx, 3ΔΔ... □ items in the column (or vertical) are 1 ◎. ◎ 2 □□, 3, ◇◇ ... "is generated, and the generated read-out data is reproduced by speech synthesis.

  The request input unit 20 is connected to an input device such as a keyboard, detects an operation of the user's input device, and inputs an individual cell reading instruction including position information of the element cell based on the detected operation. Here, in the case of a table format composed of rows and columns, the position information is indicated by a combination of row numbers and column numbers. For example, when the user inputs a keyboard such as “1”, “Enter”, “3”, and “Enter”, position information indicating one row and three columns is input. Further, when a specific keyboard key is operated by the user, the request input unit 20 inputs whole reading instruction information indicating an instruction to read the whole.

  The element output unit 22 reads out the document data from the data storage unit 32 based on the instruction information input from the request input unit 20, generates the data for reading out from the document data of the corresponding cell of the read out document data, Read-out data is reproduced by speech synthesis. At this time, when individual cell reading instruction information is input and position information including a row number and a column number is included, data for reading is generated based on document data included in a cell corresponding to the position information. Is done. In the case of a table format of only rows or columns, only one number is included as position information. Therefore, in the case of a table format of only rows, read-out data that reads out the corresponding row number from left to right is included. In the case of a table format that includes only columns, read-out data is generated to read out the corresponding column numbers from top to bottom.

  In the case where the entire reading instruction information is input to the element output unit 22 and the document data is in a tabular format consisting of rows and columns, the upper left cell excluding the row and column of items is used as a base point. Read-out data for continuously reading out from left to right and from top to bottom is generated for each line.

  In the case where the entire reading instruction information is input to the element output unit 22 and the document data is in a tabular format including only rows, the element output unit 22 starts from the next line excluding the item line. In the case of tabular document data composed of columns only, the element output unit 22 generates the next line excluding the column of items. Data for reading to be read out from the top to the bottom may be generated for each column.

  It should be noted that the order of designation of rows and columns may be designated in the order of rows and columns as described above, or may be designated in the order of columns and rows.

  Further, when generating data for reading, it may be generated so as to include only the document data of the cell including the element, or the data for reading including the document data of the cell including the item corresponding to the cell including the element to be read. May be generated.

  In addition, the instruction information input from the request input unit 20 can include information designating whether the cells divided by the integration determination / division unit 14 are individually read out or integrated and read out. When the information to be specified is included, the element output unit 22 reads the division correspondence information between the division source cell and the divided cell stored in the integrated information storage unit 30, and summarizes the information based on the read division correspondence information. It is also possible to generate read-out data to be read out as one element.

  When the document data included in the cell is a numerical value, the instruction information input from the request input unit 20 is designated to read out items having an average and maximum value and items having a minimum value for each row or column. Information can also be included.

  Next, a procedure for dividing a cell in which a plurality of cells are integrated in tabular document data described in HTML will be described with reference to FIGS.

  FIG. 2A is a diagram showing document data in a tabular format described in HTML, and FIG. 2B is a table when the document data in FIG. 2A is displayed by a web browser or the like. It is the figure which showed the structure. As shown in FIG. 2A, whether or not the cells are integrated is determined based on the value of “Colspan” included in the <TD> attribute of HTML, and whether or not the columns are combined and the number of connections. It is possible. Similarly, when “Rowspan” is included in the <TD> attribute, it is possible to determine whether or not a row is joined, and it is possible to determine the number of joins based on the value of “Rowspan”. is there.

  Next, referring to the flowchart of FIG. 3, the document data in the tabular format including the integration of a plurality of columns shown in FIG. 4A by the integration determining / dividing unit 14 is shown in the tabular format shown in FIG. A procedure for dividing the document data as described above will be described.

  First, the integration determination / division unit 14 detects a set of <TR> and </ TR> locations in order from the top in the document data in the table format detected and input by the table detection unit 12 (step Sa1). ). Next, the integration determining / dividing unit 14 detects a pair of <TD> and </ TD> in order from the top between the detected <TR> and </ TR> (step Sa2). Then, the integration determination / division unit 14 determines whether or not Colspan is specified in the detected <TD>, that is, whether or not the HTML description is “<TD Colspan = X>” (step Sa3). . If Colspan is not specified, the integration determination / division unit 14 determines whether the next combination of <TD> and </ TD> exists, and if it exists, repeats the processing from step Sa2 onward (first step). 2 loops).

  On the other hand, when the Colspan is designated, the integration determination / division unit 14 determines that the column is integrated. For example, in the example of FIG. 4A, since the cell in the first column of the second row is “<TD Colspan = 2> e </ TD>”, the column is integrated. Is determined. Next, the integration determining / dividing unit 14 sets the value indicated by “X” of Colspan = X to the variable n, and sets the document data existing between <TD> and </ TD> to the variable Value. Then, the description of “Colspan = X” is deleted. For example, in the case of the cell in the first column of the second row in FIG. 4A, the variable n is set to “2”, and the variable Value is set to “e”. Then, the location of “Colspan = 2” is deleted from “<TD Colspan = 2> e </ TD>” to become “<TD> e </ TD>” (step Sa4).

  While n is larger than 1 (step Sa5), the description of “<TD> Value </ TD>” is added after the combination of <TD> and </ TD> in which the description of “Colspan = X” is deleted. Add (step Sa6), and substitute the value of n-1 for the variable n (step Sa7). When n is 1 or less, it is determined whether or not the next combination of <TD> and </ TD> exists, and if it exists, the processing after step Sa2 is repeated (second loop).

  When the next combination of <TD> and </ TD> does not exist, the integration determination / division unit 14 determines whether or not the next combination of <TR> and </ TR> exists, and exists Repeats the processing after step Sa1 (first loop). On the other hand, if it does not exist, the process ends.

  By repeating the above processing, the document data shown in FIG. 4A is rewritten as shown in FIG. When the HTML document data of FIG. 5A is displayed by a web browser or the like, as shown in FIG. 5B, the integration of the second row and the first column and the third row integrated in FIG. 4B is performed. The divided cells are displayed in a divided state.

  If the above processing is performed for Rowspan instead of Colspan, it is possible to determine integration for rows. However, for Rowspan, when division is performed, a procedure for inserting a cell to be divided between <TR> and </ TR> is added.

  Next, operations related to the document data reading apparatus 1 will be described with reference to FIGS. First, the input unit 10 inputs document data including the table format shown in FIG. 7 (step Sb1). The table detector 12 detects a tabular part from the document data input from the input unit 10 (step Sb2). The integration determining / dividing unit 14 determines whether there is a cell in which a plurality of rows or a plurality of columns are integrated in the tabular document data (step Sb3). The integration determining / dividing unit 14 performs division when there is an integrated cell. In the case of an HTML document, division is performed according to the division procedure described above. By the division, the document data in the table format shown in FIG. 7 is divided like the document data in the table format shown in FIG. After dividing the document data including the table format, the integration determination / division unit 14 inputs the divided document data to the analysis unit 18 and also displays a division correspondence indicating the correspondence between the division source cell and the divided cell. It records in the integrated information storage part 30 (step Sb4). Note that, as described above, in the tabular document data in FIG. 7, the upper left cell includes items of diagonal lines, rows, and columns, so that the cell is not divided.

  Next, the determination unit 18a of the analysis unit 18 determines whether the upper left cell of the tabular document data input from the integration determination / division unit 14 is a blank, diagonal line, or diagonal line / row / column item. Is determined (step Sb5). If the determination unit 18a determines that the upper left cell is a blank, diagonal line, or diagonal line and row and column item, whether or not the upper left cell is a cell in which a plurality of rows or a plurality of columns are integrated. Is determined (step Sb6). When the determination unit 18a determines that the upper left cell is not a cell in which a plurality of rows or columns are integrated, a cell including items and a cell including elements are specified as tabular text data in one row and one column. . Then, the determination unit 18a inputs information specifying the cell including the item to the item number calculation unit 18b, and inputs information specifying the cell including the element to the element number calculation unit 18c. Information specifying the cell including the element and document data are input to the overall output unit 24 and the item output unit 26 (step Sb7). Here, the specified information is information in which identification information indicating an item or element is associated with position information indicating the position of each cell.

  On the other hand, in step Sb6, when the determination unit 18a determines that the upper left cell is a cell in which a plurality of rows or a plurality of columns are integrated, the number of rows from the top is an item of the row, and the number of columns from the left is the column. Specify whether it is an item, and specify the cell that contains the item. Then, the determination unit 18a inputs information specifying the cell including the item to the item number calculation unit 18b, and inputs information specifying the cell including the element to the element number calculation unit 18c. Information specifying the cell including the element and document data are input to the overall output unit 24 and the item output unit 26 (step Sb8).

  On the other hand, when the determination unit 18a determines in step Sb5 that the upper left cell is not a blank, diagonal line, or diagonal line and row and column item, for example, tabular document data detected by the table detection unit 12 9 and FIG. 10, the determination unit 18a determines whether the tabular document data is only a row or only a column (step Sb9).

  Here, in order to determine whether the tabular document data is only a row or only a column, the contents of vertical cells such as the contents of cells in the first row and the contents of cells in the second and subsequent rows are used. When the contents are sequentially traced and the contents of the cells in the first, second, and third columns are compared, the contents of the cell in the first row and the contents of the cells in the second and subsequent rows, or one row When there is a tendency to divide into two, such as the contents of the cells in the first and second rows and the contents of the cells in the third and subsequent rows, there is no item in the column, and the document data is in tabular format with only rows Judge that there is.

  On the other hand, if it is not possible to detect the tendency that the data format is divided into two even if the contents of the vertical cell are traced, the contents of the horizontal cells such as the contents of the cells in the first column and the contents of the cells in the second column and after are If the contents of the cells in the first row, the second row, and the third row are compared and the vertical cells are traced, the items in the row are It is determined that the document data is in a tabular format with only columns.

  In the case of a table format consisting of only rows or columns by the above processing, in the case of document data that does not lead to an understanding of the element information expressed by the table, based on the row and column items of each cell, For example, in the case of tabular document data in which the element information is clarified by representing the increase / decrease or change of the value, such as each month in the first column and the transition of each month in the second column. Even if it exists, it can specify that it is the table format which consists only of a row or a column.

  For example, when applied to the document data in the table format shown in FIG. 9, even if the contents of the vertical cells are traced, the tendency that the data format is divided into two cannot be detected, and the contents of the horizontal cells are sequentially changed. Traced, the combination of the first row and the second row is the ward name and the town name, the cell in the third row is the name of the place, and is divided into two, so it is document data in a tabular format only in columns. It is determined.

  Further, when applied to the tabular document data shown in FIG. 10, when the contents of the vertical cells are traced, the first column is the day of the week, whereas the second and subsequent columns are garbage types. It is determined that the document data is in a tabular format with only rows.

  The determination of the data format is performed by determining whether or not the items included in the cell are natural language and numbers, and the data format between the natural languages is stored in advance in advance. This is done by determining the type of category to which the language belongs, such as the name of a car or the type of garbage, based on the dictionary data.

  When the determination unit 18a determines that the tabular document data is only a row, in the above-described determination, the cell including the item and the cell including the element are specified using the divided line as a base point. Then, the determination unit 18a inputs information specifying the cell including the item to the item number calculation unit 18b, and inputs information specifying the cell including the element to the element number calculation unit 18c. Information specifying the cell including the element and document data are input to the overall output unit 24 and the item output unit 26 (step Sb10). On the other hand, when the determination unit 18a determines that the tabular document data is only a column, in the above-described determination, the cell including the item and the cell including the element are specified based on the divided column. . Then, the determination unit 18a inputs information specifying the cell including the item to the item number calculation unit 18b, and inputs information specifying the cell including the element to the element number calculation unit 18c. Information specifying the cell including the element and document data are input to the overall output unit 24 and the item output unit 26 (step Sb11).

  And the determination means 18a specified the cell containing an item, and the cell containing an element, when the cell containing an item and the cell containing an element were specified in step Sb7, step Sb8, step Sb10, and step Sb11. Information and document data are recorded in the data storage unit 32 (step Sb12).

  Next, the item number calculation unit 18b counts the number of cells including the item based on the information specifying the cell including the item input from the determination unit 18a, and outputs the counted value to the overall output unit 24. (Step Sb13). In addition, the element number calculation unit 18c counts the number of cells including the element based on the information specifying the cell including the element input from the determination unit 18a, and outputs the counted value to the overall output unit 24. (Step Sb14).

  The total output unit 24 includes information and document data specifying whether the cell input from the determination unit 18a includes an item or an element, and the value of the number of cells including the item input from the item number calculation unit 18b. Then, based on the value of the number of cells including the elements input from the element number calculation means 18c, reading data indicating the entire document data in the tabular format is generated (step Sb15), and the generated reading data is Playback is performed by speech synthesis (step Sb16). The item output unit 26 generates read-out data indicating the contents of each item based on the specified information and document data input from the determination unit 18a, and reproduces the generated read-out data by speech synthesis ( Step Sb17). After the read-out data is reproduced by speech synthesis from the overall output unit 24 and the item output unit 26, the user operates the input device connected to the request input unit 20, and the request input unit 20 that has received the operation Instruction information for reading a cell including an element is input to the element output unit 22, and the element output unit 22 reads out document data from the data storage unit 32 based on the input instruction information, and reads out the document data from the read document data. Data for reading is generated, and the generated data for reading is reproduced by voice synthesis (step Sb18).

  Here, the processing from step Sb5 to step Sb12 will be described below by taking the tabular document data after division shown in FIG. 8 as an example. First, the determination unit 18a determines that the upper left cell is an item of diagonal lines, rows, and columns based on the document data of FIG. 8, and the upper left cell is a cell in which a plurality of rows are integrated. Determine. That is, the tabular document data is a cell including items from the top to the second row and from the left to the first column, excluding the upper left cell, and the cells from the third row and the second column. Identifies a cell that contains an element. Then, the item number calculating means 18b counts the cells including the items, and outputs the number of cells as four rows and ten columns. The element number calculating means 18c counts the number of cells including the element and calculates that the number is 20.

  The overall output unit 24 calculates information and document data specifying whether to include an item or an element input from the determination unit 18a, and the number of cells including an item input from the item number calculation unit 18b and the number of elements. Based on the number of cells that contain the elements input from the means 18c, “There are 5 districts as horizontal items in the 2 rows from the top. 4 days of the week as vertical items in 1 column from the left. "There are a total of 20 data in 4 rows and 5 columns." At this time, the text data explaining the vertical and horizontal items using the row and column items of the upper left cell is also included in the reading data. Then, the overall output unit 24 reproduces the generated reading data by voice synthesis.

  The item output unit 26 uses the information specifying whether the cell includes an item or an element and the document data as follows: “The five districts indicated on the side are 1, Higashi-ku, A-machi, 2, East. Ward, B Town, 3, Nishi Ward, C Town, 4, Nishi Ward, D Town, 5, Nishi Ward, E Town, etc. The four days of the week listed are 1, Monday, 2, Wednesday, 3, Thursday. , 4, Friday "is generated. At this time, the text data explaining the vertical and horizontal items using the row and column items of the upper left cell is also included in the reading data. Then, the item output unit 26 reproduces the generated read-out data by voice synthesis.

  Further, the process of step Sb13 will be described below using the document data of FIG. 8 as an example. First, the case where only necessary elements are read out will be described. The request input unit 20 is preset to be input in the order of columns and rows, and the element output unit 22 includes text data of a cell including the specified element and items corresponding to the element. It is assumed that the document data column and row are set to be read.

  First, when the user operates the input device and inputs “3”, “Enter”, “2”, and “Enter”, the request input unit 20 outputs the individual cell reading instruction information including the position information of the third column and the second row as an element. Input to the unit 22. When the individual cell reading instruction information is input, the element output unit 22 is included in the cell of the third column and the second row from the data storage unit 32 based on the position information indicating the third column and the second row included in the instruction information. Read the document data “combustible waste”, read the document data “Nishi-ku” and “C town” included in the item cell of the third column, read the document data “Wednesday” included in the item cell of the second row, On Wednesday of Nishi Ward and C town, the data for reading “combustible waste” is generated, and the generated reading data is reproduced by voice synthesis.

  In addition, when the element output unit 22 generates data for reading “combustible waste on Wednesday in Nishi Ward, C town” from document data “burnable waste”, “Nishi Ward”, “C town”, “Wednesday”, The particle "-no" and "-ha" which are not included in these document data are stored in advance as information on particles used to connect the document data for each item and cell relationship. The information of the particle stored in the output unit 22 is read out and added between the document data based on the relationship between the item and the cell to generate the above-mentioned document data “Nishi-ku, Wednesday in C town combustible waste”.

  Contrary to the above example, the request input unit 20 is preset to be input in the order of rows and columns, and the element output unit 22 is the text data of the cell including the specified element. It is assumed that the document data of the item corresponding to the element is set to be read in the order of rows and columns. When the user operates the input device and inputs “3”, “Enter”, “3”, and “Enter”, the request input unit 20 outputs the individual cell reading instruction information including the position information of the third row and third column as an element Input to the unit 22. When the individual cell reading instruction information is input, the element output unit 22 is included in the cell of the third row and the third column from the data storage unit 32 based on the position information indicating the third row and the third column included in the instruction information. Read the document data “bottle / can / pet bottle”, read the document data “Thursday” contained in the item cell in the third row, and read the document data “Nishi-ku” and “C town” contained in the item cell in the third column. Is read out, “Thursday, Nishi-ku, C town is bottle / can / pet bottle” is generated, and the generated read-out data is reproduced by speech synthesis.

  Next, a case where the entire table is read out will be described. When the user operates the input device to input information for designating whole reading, the request input unit 20 inputs whole reading instruction information to the element output unit 22. The element output unit 22 reads out the document data from the data storage unit 32 and reads “combustible garbage on Mondays in East and A towns, combustible garbage on Mondays in East and B towns, and non-burnable garbage on Mondays in Nishi and C towns”. ,..., Read-out data to be read out from left to right in order from the first line is generated, and the generated read-out data is reproduced by speech synthesis.

  If the entire reading instruction information includes an instruction to read the divided cells together, the element output unit 22 refers to the integrated information storage unit 30 and displays the tabular document data before division shown in FIG. Based on the composition of "While in Higashi Ward, A town, and B town Monday is combustible waste, Monday in both Nishi Ward, C town, and D town, Monday in Nishi Ward, E town, combustible waste, Higashi Ward, A town" From Wednesday to Wednesday, incombustible waste, Higashi Ward, B town Wednesdays, bottles, cans, PET bottles, Nishi Ward, C town, D town, Wednesday combustible waste, etc. Read-out data to be read out is generated, and the generated read-out data is reproduced by speech synthesis.

  With the configuration of the above-described embodiment, the overall output unit 24 reads out the entire tabular configuration, and further, the item output unit 26 reads out the contents of each item, so that it is difficult to use visually. This makes it possible to understand the overview of the table and the summary of items that are supposed to be captured by the eye, and has the effect of facilitating the understanding of tabular document data for visually impaired persons.

  The input unit of the present invention corresponds to the input unit 10, and the table detection unit corresponds to the table detection unit 12. The determination unit, the item number calculation unit, and the element number calculation unit of the present invention correspond to the determination unit 18a, the item number calculation unit 18b, and the element number calculation unit 18c of the analysis unit 18. The integrated determination / division means of the present invention corresponds to the integrated determination / division unit 14. In the present invention, the request input means corresponds to the request input section 20, the element output means corresponds to the element output section 22, the overall output means corresponds to the overall output section 24, and the item output means corresponds to the item output section 26.

  The document data reading apparatus 1 described above has a computer system therein. The integrated cell dividing process and the process of reproducing tabular document data by speech synthesis are stored in a computer-readable recording medium in the form of a program, which is read by the computer. By executing this, the above processing is performed. Here, the computer-readable recording medium means a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, a semiconductor memory, or the like. Alternatively, the computer program may be distributed to the computer via a communication line, and the computer that has received the distribution may execute the program.

It is a block diagram which shows the internal structure of the document data reading apparatus which concerns on this embodiment. It is the figure which showed the document data of the table format described by HTML concerning the embodiment. It is the flowchart which showed the division | segmentation process of the cell by which the document data of the table format described by HTML concerning the embodiment were integrated. It is the figure which showed the state before the division | segmentation of the document data of the table format described by HTML concerning the embodiment. It is the figure which showed the state after the division | segmentation of the document data of the table format described by HTML concerning the embodiment. It is the flowchart which showed the reading-out process of the document data containing the table format based on the embodiment. It is the figure which showed the document data (before division | segmentation) of the table format input into the document data reading apparatus which concerns on the embodiment. It is the figure which showed the document data (after division | segmentation) of the table format input into the document data reading apparatus which concerns on the embodiment. It is the figure which showed the document data of the table format which consists only of the column which concerns on the embodiment. It is the figure which showed the document data of the table format which consists only of the line which concerns on the embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Document data reading apparatus 10 Input part 12 Table | surface detection part 14 Integrated determination / division | segmentation part 18 Analysis part 18a Determination means 18b Item number calculation means 18c Element number calculation means 20 Request input part 22 Element output part 24 Whole output part 26 Item output part 30 integrated information storage unit 32 data storage unit

Claims (5)

A text data reading device that reads and outputs input text data by speech synthesis,
Sentence data including table information indicating a correspondence between a position of a cell in a table including one or more rows and one or more columns and an item included in the cell or an element of the item included in the cell. Input means for inputting;
Table detection means for detecting the table information from sentence data input by the input means;
Determining means for determining whether each cell of the table information detected by the table detecting means is a cell including the item or a cell including the element;
Item number calculating means for calculating the number of cells determined by the determining means as cells containing items;
Element number calculating means for calculating the number of cells determined by the determining means as cells containing elements;
Whole output means for reading out and outputting the number of cells calculated by the item number calculating means as the number of items by speech synthesis and outputting the number of cells calculated by the element number calculating means as the number of elements by speech synthesis When,
Item output means for reading out and outputting text data contained in all cells determined by the determination means as cells containing items by speech synthesis;
Element output means for reading out and outputting text data included in the cell determined by the determination means as a cell including an element by speech synthesis;
A text data reading device comprising: The table information detected by the table detection means determines whether or not an integrated cell is included, and if it is determined that an integrated cell is included, the integrated cell is divided, and the original cell is divided Including an integrated determination division means for outputting table information in which the contents of items or elements included in the contents of all divided cells are output,
The determination means includes
When the integrated determination dividing unit determines that the cell does not include an integrated cell, it is determined whether each cell of the table information detected by the table detection unit is a cell including the item or a cell including the element. When it is determined that the integrated determination dividing unit includes an integrated cell, each cell of the table information output from the integrated determination dividing unit is a cell including the item or a cell including the element. The sentence data reading device according to claim 1, wherein the text data is read out. In response to a user operation, a request input unit is provided for inputting position information for designating a position of a cell including the element among the cells in the table format.
The element output means includes
The text data reading device according to claim 1 or 2, wherein text data included in a cell corresponding to position information input by the request input means is read out by voice synthesis and output. The determination means includes
It is determined whether the table information is a table having items in rows and columns , a table having items only in rows, or a table having items only in columns , and a table having items in rows and columns. when determining whether the cells of the table information is a cell that contains the entry corresponding to the row, or cell containing the item corresponding to the column, to determine whether the cell containing the element, the row only If it is determined that a table with an entry, or each cell of the table information is a cell that contains the entry corresponding to the line, to determine whether the cell containing the element, determining a table with fields only in the column In this case, it is determined whether each cell of the table information is a cell including the item corresponding to a column or a cell including the element. The text data reading device described. A computer of a text data reading device that reads and outputs input text data by speech synthesis,
Sentence data including table information indicating a correspondence between a position of a cell in a table including one or more rows and one or more columns and an item included in the cell or an element of the item included in the cell. Input means for inputting;
Table detection means for detecting the table information from sentence data input by the input means;
Determining means for determining whether each cell of the table information detected by the table detecting means is a cell including the item or a cell including the element;
Item number calculating means for calculating the number of cells determined by the determining means as cells containing items;
Element number calculating means for calculating the number of cells determined by the determining means as cells containing elements;
Whole output means for reading out and outputting the number of cells calculated by the item number calculating means as the number of items by speech synthesis and outputting the number of cells calculated by the element number calculating means as the number of elements by speech synthesis When,
Item output means for reading out and outputting text data contained in all cells determined by the determination means as cells containing items by speech synthesis;
Element output means for reading out and outputting text data included in the cell determined by the determination means as a cell including an element by speech synthesis;
Computer program to function as.

Images