Document reading apparatus, method of controlling the same, and storage medium

Abstract

To provide a business form reading device, a control method thereof, and a program capable of previously grasping table formats of tables of various business forms to be read.SOLUTION: In the form reading device including a reading unit, an extraction unit, a storage unit, a printing unit, an operation unit, a display unit, and a control unit, the control unit detects a first degree of similarity between a table of a form read by the reading unit and a reference table pattern in the storage unit, and when the detected first degree of similarity is less than a S3, the control unit determines that the reference table pattern includes, as a table, a portion corresponding to the table of the form read by the reading unit in the S4NO. Generating a reference table pattern in a new table format that does not include other portions as a table (S5), detecting a second similarity between the table of the form read by the reading unit and the generated reference table pattern in the new table format (S6), and storing the generated reference table pattern in the new table format in the storage unit (S7YES) when the detected second similarity is equal to or more than the predetermined value (S10).SELECTED DRAWING: Figure 7

Description

[Technical Field]

[0001] An embodiment of the present invention relates to a document reading device that reads documents that include tables, a control method for the document reading device, and a program. [Background technology]

[0002] In a document reading device that reads documents that include tables, in order to quickly and efficiently recognize the tables on the scanned documents and their contents, it is necessary to know in advance the table format of the tables on the various documents to be read. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2016-081380 Summary of the Invention [Problem to be solved by the invention]

[0004] An object of the embodiment is to provide a document reading device that can grasp in advance the table format of tables in various documents to be read, as well as a control method and program for the same. [Means for solving the problem]

[0005] A form reading device according to an embodiment includes a reading unit that reads a form including a table; a storage unit that stores a predetermined reference table pattern in a tabular format; and a control unit that recognizes the table of the form read by the reading unit and the information in the table based on the reference table pattern in the storage unit. The control unit includes: first detection means that detects a first similarity between the table of the form read by the reading unit and the reference table pattern in the storage unit; generation means that, if the first similarity detected by the detection means is less than a threshold, generates a new reference table pattern in a tabular format that includes, as a table, a portion of the reference table pattern that corresponds to the table of the form read by the reading unit and does not include other portions as tables; second detection means that detects a second similarity between the table of the form read by the reading unit and the new reference table pattern in a tabular format generated by the generation means; and control means that, if the second similarity detected by the second detection means is equal to or greater than the threshold, stores the new reference table pattern in a tabular format generated by the generation means in the storage unit. [Brief explanation of the drawings]

[0006] [Figure 1] FIG. 1 is a block diagram showing the configuration of one embodiment. [Figure 2] FIG. 2 is a diagram showing a standard form in the embodiment. [Figure 3] FIG. 3 is a diagram showing an example of a document to be read in the embodiment. [Figure 4] FIG. 4 is a diagram showing another example of a document to be read in the embodiment. [Figure 5] FIG. 5 is a diagram showing another example of a document to be read in the embodiment. [Figure 6] FIG. 6 is a diagram showing yet another example of a document to be read in the embodiment. [Figure 7] FIG. 7 is a flowchart showing the control of the embodiment. [Figure 8] FIG. 8 is a diagram showing how a new reference table pattern in a table format is generated in the embodiment. [Figure 9] FIG. 9 is a diagram showing a new reference table pattern in a tabular format generated in the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0007] A document reading device according to an embodiment will be described with reference to the drawings. As shown in FIG. 1, the document reading device includes a reading unit 1, an extraction unit 2, a storage unit 3, a printing unit 4, an operation unit 5, a display unit 6, and a control unit 10.

[0008] The reading unit 1 is an imaging device such as a camera that optically reads the images of multiple forms 30, 40, 50, 60... that are set as reading targets, one by one. Each of these forms 30, 40, 50, 60... includes a table consisting of one or more character recording lines. The extraction unit 2 extracts the table pattern, which is a characteristic part of the forms 30, 40, 50, 60... that are read by the reading unit 1.

[0009] The storage unit 3 stores (registers) in advance a reference table pattern 21 having a predetermined table format, for example, a table format with a fixed number of rows. The reference table pattern 21 is included in a reference form 20 shown in FIG.

[0010] The control unit 10 is configured by a microcomputer and its peripheral circuits, and recognizes the table extracted by the extraction unit 2 and the information in the table based on the reference table pattern 21 (and the reference table pattern 71) in the storage unit 3.

[0011] [Standard Report 20 (Fixed Row Table Format)] The reference form 20 is a rectangular piece of paper consisting of a top edge, a left edge, a right edge, and a bottom edge, as shown in Figure 2, and includes a rectangular reference table pattern 21 at a predetermined distance from each of the top edge, left edge, right edge, and bottom edge.

[0012] The reference table pattern 21 is formed by a plurality of parallel vertical lines 21a extending vertically along the left and right sides of the reference form 20 and a plurality of parallel horizontal lines 21b extending horizontally along the top and bottom sides of the reference form 20, and has a table format with a fixed number of rows including a plurality of character recording rows 21c, for example, eight rows, between each horizontal line 21b.

[0013] Each character record line 21c includes a plurality of character record frames 21d, for example, 24. The title (name) of the table 21 is written in the space between the top edge of the table 21 and the top edge of the base form 20.

[0014] [Report 30 (table format with fixed number of rows)] As shown in FIG. 3, the document 30 is a rectangular piece of paper consisting of a top edge, a left edge, a right edge, and a bottom edge, and includes a rectangular table pattern 31 at a predetermined distance from each of the top edge, the left edge, the right edge, and the bottom edge.

[0015] The table pattern 31 is formed by a plurality of parallel vertical lines 31a extending vertically along the left and right sides of the form 30 and a plurality of parallel horizontal lines 31b extending horizontally along the top and bottom sides of the form 30, and includes a plurality of character recording lines 31c, for example, eight lines, between each horizontal line 31b.

[0016] The eight character record rows 31c include a plurality of character record frames 31d, for example, 24 character record frames 31d. Character information such as kanji, numbers, and alphabetic characters is recorded in all of these character record frames 31d. The title (name) 31e of table 31 is written in the space between the top edge of table 31 and the top edge of form 30.

[0017] The number of lines "8" in each character recording line 31c is the same as the number of lines "8" in the character recording line 21c in the reference table pattern 21 of the reference form 20. In other words, the table format of the form 30 is the same fixed-number-of-lines type as the table format of the reference form 20. The number of character recording frames 31d, "24", is also the same as the number of character recording lines 21c in the reference table pattern 21 of the reference form 20.

[0018] [Report 40 (table format with fixed number of rows)] As shown in FIG. 4, the document 40 is a rectangular piece of paper consisting of a top edge, a left edge, a right edge, and a bottom edge, and includes a rectangular table pattern 41 at a predetermined distance from each of the top edge, the left edge, the right edge, and the bottom edge.

[0019] The table pattern 41 is formed by a plurality of parallel vertical lines 41a extending vertically along the left and right sides of the form 40 and a plurality of parallel horizontal lines 41b extending horizontally along the top and bottom sides of the form 40, and includes a plurality of character recording lines 41c, for example, eight lines, between each horizontal line 41b.

[0020] The eight character record lines 41c include a plurality of character record frames 41d, for example, 24. The title (name) 41e of the table 41 is written in the space between the top edge of the table 41 and the top edge of the form 40.

[0021] The number of lines "8" in each character recording line 41c is the same as the number of lines "8" in the character recording lines 21c in the reference table pattern 21 of the reference form 20. In other words, the table format of the form 40 is the same fixed-number-of-lines type as the table format of the reference form 20. The number of character recording frames 41d, "24", is also the same as the number of character recording lines 21c in the reference table pattern 21 of the reference form 20.

[0022] Character information such as kanji characters, numbers, and alphabetic characters is recorded in all character recording frames 41d of the first, second, and third character recording lines 41c from the top, and no character information is recorded in any of the character recording frames 41d of the remaining fourth to eighth character recording lines 41c, leaving them blank.

[0023] [Report 50 (variable row count table format)] As shown in FIG. 5, the document 50 is a rectangular piece of paper consisting of a top edge, a left edge, a right edge, and a bottom edge, and includes a rectangular table pattern 51 at a predetermined distance from each of the top edge, the left edge, the right edge, and the bottom edge.

[0024] The table pattern 51 is formed by a plurality of parallel vertical lines 51a extending vertically along the left and right sides of the form 50 and a plurality of parallel horizontal lines 51b extending horizontally along the top and bottom sides of the form 50, and includes a plurality of character recording lines 51c, for example, three lines, between each horizontal line 51b.

[0025] The three character recording lines 51c include a plurality of character recording boxes 51d, for example, nine. Character information such as kanji characters, numbers, and alphabetic characters is recorded in all of the character recording boxes 51d of the three character recording lines 51c. The title (name) 51e of the table 51 is written between the top edge of the basic form 50 and the top edge of the table 51.

[0026] The number of lines "3" in each character recording line 51c is different from the number of lines "8" in the character recording lines 21c in the reference table pattern 21 of the reference form 20. In other words, the table format of the form 50 is a variable-line type, which is different from the table format (fixed-line type) of the reference form 20. The number of character recording frames 51d, "9," is also different from the number of character recording lines 21c, "24," in the reference table pattern 21 of the reference form 20.

[0027] The size (area) of table pattern 51 having three character recording lines 51c is smaller than the size (area) of tables 31 and 41 having eight character recording lines 31c and 41c like forms 30 and 40. The area below table pattern 51 in form 50 is a blank area 51f with no entries.

[0028] It may be difficult to determine whether this blank area 51f really has no writing in it or is a blank space like the character recording lines 41c on lines 4 to 8 of table 41 of document 40, simply by the reading process of reading unit 1.

[0029] [Report 60 (variable row count table format)] As shown in FIG. 6, the document 60 is a rectangular piece of paper consisting of a top edge, a left edge, a right edge, and a bottom edge, and includes an approximately rectangular table pattern 61 at a predetermined distance from each of the top edge, the left edge, the right edge, and the bottom edge.

[0030] The table pattern 61 is formed by a plurality of parallel vertical lines 61a extending vertically along the left and right sides of the form 50 and a plurality of parallel horizontal lines 61b extending horizontally along the top and bottom sides of the form 60, and includes a plurality of character recording lines 61c, for example, three lines, between each horizontal line 61b.

[0031] Each of the three character recording lines 61c includes a plurality of character recording frames 61d, for example, eight. In particular, the number of character recording frames 61d in the third (last) character recording line 61c is two, one less than the number of character recording frames 61d in the first and second character recording lines 61c, which is three. Therefore, the table pattern 61 has a stepped shape with a notch 61g in the lower left corner. The title (name) 61e of the table 61 is written in the space between the top edge of the table 61 and the top edge of the form 60.

[0032] The number of lines "3" in each character recording line 61c is different from the number of lines "8" in the character recording lines 21c in the reference table pattern 21 of the reference form 20. In this regard, the table format of the form 60 can be considered to be a variable-number-of-lines type, which is different from the table format (fixed-number-of-lines type) of the reference form 20.

[0033] The size (area) of the table pattern 61 having three character recording lines 61c is smaller than the size (area) of the tables 31 and 41 having eight character recording lines 31c and 41c like the forms 30 and 40. The area below the table pattern 61 in the form 60 is a blank area 61f with no entries.

[0034] It may be difficult to determine whether this blank area 61f really has no writing in it or is a blank space like the character recording lines 41c on lines 4 to 8 of table 41 of document 40, simply by the reading process of reading unit 1.

[0035] [Control unit 10] The control unit 10 includes, as main functions realized by executing a program stored in the internal memory, a first detection means 10a, a first determination means 10b, a control means 10c, a generation means 10d, a second detection means 10e, a second determination means 10f, and control means 10g to 10j.

[0036] The first detection means 10a detects the similarity (first similarity) X1 between the table extracted by the extraction unit 2 and the reference table pattern 21 in the storage unit 3 by calculating the following equation. X1=Xa / Xb Xa is the number (number of frames) of character recording frames in the form extracted by the extraction unit 2. Xb is the number (number of frames) of character recording frames 21d in the reference form pattern 21.

[0037] The first determination means 10b determines whether the similarity X1 detected by the first detection means 10a is equal to or greater than the threshold value Xs. The threshold value Xs is, for example, 90%.

[0038] When the determination result of the first determination means 10b is affirmative (X1≥Xs), the first control means 10c proceeds to the reading process of the reading unit 1 for the next form on the assumption that the form of the form read by the reading unit 1 is the same fixed number of lines type as the form of the reference form pattern 21 in the storage unit 3.

[0039] When the determination result of the first determination means 10b is negative (X1<Xs), the generation means 10d generates a new reference form pattern of a form type (variable number of lines type) that includes, as a form, the portion corresponding to the form extracted by the extraction unit 2 in the reference form pattern 21 and does not include the other portions as forms.

[0040] The second detection means 10e detects the similarity (second similarity) X2 between the form extracted by the extraction unit 2 and the new reference form pattern of the form type generated by the generation means 10d by the calculation of the following formula. X2 = Xa / Xc Xa is the number of character recording frames in the form extracted by the extraction unit 2. Xc is the number of character recording frames in the new reference form pattern of the form type (variable number of lines type) generated by the generation means 10d.

[0041] <0CO0171>The second determination means 10f determines whether the similarity X2 detected by the second detection means 10e is equal to or greater than the above threshold value Xs.

[0042] If the judgment result of the second judgment means 10f is positive (X2≧Xs), the control means 10g displays the new tabular reference table pattern generated by the generation means 10d, together with the table extracted by the extraction unit 2, on the display unit 6 for operator approval.

[0043] When the operator inputs information by operating the operation unit 5 indicating approval of the new tabular reference table pattern displayed on the display unit 6 by the control means 10g, the control means 10h additionally stores (registers) the new tabular reference table pattern generated by the generation means 10d in the memory unit 3, and then proceeds to the reading process of the reading unit 1 for the next document.

[0044] When the operator inputs information by operating the operation unit 5 that the new tabular reference table pattern displayed on the display unit 6 by the control means 10g is not approved, the control means 10i notifies the operator of the information by displaying it on the display unit 6 or by printing it out on the printing unit 4, and then the reading unit 1 proceeds to read the next document.

[0045] If the determination result of the second determination means 10f is negative, the control means 10j notifies the user of this information by displaying it on the display unit 6 or printing it out on the printing unit 4, and then causes the reading unit 1 to proceed with the reading process for the next document.

[0046] Next, the control executed by the control unit 10 will be described with reference to the flowchart of FIG. [Reading Form 30] An operator sets multiple forms in the reading unit 1 and issues a command to start processing using the operation unit 5. Based on this command, the reading unit 1 optically reads one of the set forms, for example, form 30 in Fig. 3 (S1). Next, the extraction unit 2 extracts a front pattern 31, which is a characteristic part of the form 30 read by the reading unit 1 (S2).

[0047] The control unit 10 calculates the similarity X1 between the table pattern 31 extracted by the extraction unit 2 and the reference table pattern 21 of FIG. 2 stored in the storage unit 3 by the following formula (S3). X1=Xa / Xb=24 / 24=100% Xa is the number (=24) of character recording frames 31d in the table pattern 31 extracted by the extraction unit 2. Xb is the number (=24) of character recording frames 21d in the reference table pattern 21.

[0048] Then, the control unit 10 determines whether the detected similarity X1 is equal to or greater than a threshold value Xs (for example, 90%) (S4). In this case, since the similarity X1 is greater than or equal to the threshold value Xs (X1≧Xs), the control unit 10 determines that the table format of the document 30 read by the reading unit 1 is the same fixed row number type as the table format of the reference document 20, and proceeds to the reading process of the reading unit 1 for the next document, for example, document 40 in Figure 4 (S1).

[0049] [Reading Form 40] The extraction unit 2 extracts a table pattern 41, which is a characteristic part of the document 40 read by the reading unit 1 (S2). Then, the control unit 10 calculates the similarity X1 between the table pattern 41 extracted by the extraction unit 2 and the reference table pattern 21 of Fig. 2 in the memory unit 3 by the following equation (S3). X1=Xa / Xb=24 / 24=100% Xa is the number (=24) of character recording frames 41d in the table pattern 41 extracted by the extraction unit 2. Xb is the number (=24) of character recording frames 21d in the reference table pattern 21.

[0050] Then, the control unit 10 determines whether the detected similarity X1 is equal to or greater than a threshold value Xs (S4). In this case, since the similarity X1 is greater than or equal to the threshold value Xs (X1≧Xs), the control unit 10 determines that the table format of the document 40 read by the reading unit 1 is the same fixed row number type as the table format of the reference document 20, and proceeds to the reading process of the reading unit 1 for the next document, for example, document 50 in Figure 5 (S1).

[0051] [Reading Form 50] The extraction unit 2 extracts a table pattern 51, which is a characteristic part of the document 50 read by the reading unit 1 (S2). Then, the control unit 10 calculates the similarity X1 between the table pattern 51 extracted by the extraction unit 2 and the reference table pattern 21 of Fig. 2 in the memory unit 3 by the following equation (S3). X1=Xa / Xb=9 / 24=37.5% Xa is the number (=9) of character recording frames 51d in the table pattern 51 extracted by the extraction unit 2. Xb is the number (=24) of character recording frames 21d in the reference table pattern 21.

[0052] Then, the control unit 10 determines whether the detected similarity X1 is equal to or greater than a threshold value Xs (S4). In this case, since the similarity X1 is less than the threshold value Xs (NO in S4), the control unit 10 generates a new reference table pattern of variable row number type (S5) based on the determination that the table format of the form 50 read by the reading unit 1 is a variable row number type different from the table format of the reference form 20. That is, as shown in Fig. 8, the control unit 10 generates a new reference table pattern of table format (variable row number type) that includes, as a table, the portion of the reference table pattern 21 that corresponds to the table pattern 51 extracted by the extraction unit 2 (the portion indicated by the solid line in the figure) and does not include the other portion (the portion indicated by the dashed line in the figure) as a table.

[0053] The new table format (variable number of rows) reference table pattern thus generated is table pattern 71 of form 70 as shown in Figure 9. That is, form 70 is a rectangular piece of paper consisting of a top side, a left side, a right side, and a bottom side, and includes rectangular table patterns 71 at positions at a predetermined distance from each of the top side, the left side, the right side, and the bottom side. Hereinafter, table pattern 71 will be referred to as new table format reference table pattern 71.

[0054] The new tabular reference table pattern 71 is formed by a plurality of parallel vertical lines 71a extending vertically along the left and right sides of the form 70 and a plurality of parallel horizontal lines 71b extending horizontally along the top and bottom sides of the reference form 70, with three character recording lines 71c between each horizontal line 71b and nine character recording frames 71d within each character recording line 71c.

[0055] Next, the control unit 10 calculates the similarity X2 between the table pattern 51 extracted by the extraction unit 2 and the newly generated reference table pattern 71 in table format by the following equation (S6). X2=Xa / Xc=9 / 9=100% Xa is the number (=9) of character recording frames 51d in the table pattern 51 extracted by the extraction unit 2. Xc is the number (=9) of character recording frames 71d in the new reference table pattern 71 in table format.

[0056] Then, the control unit 10 determines whether the detected similarity X2 is equal to or greater than the threshold value Xs (S7). In this case, since the similarity X2 is greater than or equal to the threshold value Xs (X2≧Xs; YES in S7), the control unit 10 displays the new tabular reference table pattern 71 on the display unit 6 for operator approval together with the table pattern 51 extracted by the extraction unit 2 (S8).

[0057] The new tabular reference table pattern 71 displayed on the display unit 6 is almost identical to the table pattern 51 displayed on the display unit 6. Because the new tabular reference table pattern 71 is almost identical to the table pattern 51, the operator can approve the new tabular reference table pattern 71 as information necessary for the reading process of the document reading device.

[0058] When the operator inputs OK information indicating approval of the new tabular reference table pattern 71 displayed on the display unit 6 by operating the operation unit 5 (YES in S9), the control unit 10 additionally registers (stores) the generated new tabular reference table pattern 71 in the storage unit 3 by adding it to the fixed-number-of-rows reference table pattern 21 (S10). Subsequently, the control unit 10 proceeds to the reading process of the reading unit 1 for the next form (S1).

[0059] [Reading Form 60] The extraction unit 2 extracts a table pattern 61, which is a characteristic part of the document 60 read by the reading unit 1 (S2). Then, the control unit 10 calculates the similarity X1 between the table pattern 61 extracted by the extraction unit 2 and the reference table pattern 21 of Fig. 2 in the memory unit 3 by the following equation (S3). X1=Xa / Xb=8 / 24=33.3% Xa is the number (=8) of character recording frames 61d in the table pattern 61 extracted by the extraction unit 2. Xb is the number (=24) of character recording frames 21d in the reference table pattern 21.

[0060] Then, the control unit 10 determines whether the detected similarity X1 is equal to or greater than a threshold value Xs (S4). In this case, since the similarity X1 is less than the threshold value Xs (NO in S4), the control unit 10 generates a new reference table pattern of variable row number type (S5) based on the determination that the table format of the form 60 read by the reading unit 1 is a variable row number type different from the table format of the reference form 20. That is, as shown in Fig. 8, the control unit 10 generates a new reference table pattern of table format (variable row number type) that includes, as a table, the portion of the reference table pattern 21 that corresponds to the table pattern 61 extracted by the extraction unit 2 (the portion indicated by the solid line in the figure), but does not include the other portion (the portion indicated by the dashed line in the figure) as a table.

[0061] The new table format (variable row count type) reference table pattern thus generated is a table pattern 71 of a form 70 as shown in FIG.

[0062] Next, the control unit 10 calculates the similarity X2 between the table pattern 61 extracted by the extraction unit 2 and the newly generated reference table pattern 71 in table format by the following equation (S6). X2=Xa / Xc=8 / 9=88.8% Xa is the number (=8) of character recording frames 61d in the table pattern 61 extracted by the extraction unit 2. Xc is the number (=9) of character recording frames 71d in the new reference table pattern 71 in table format.

[0063] Then, the control unit 10 determines whether the detected similarity X2 is equal to or greater than the threshold value Xs (S7). In this case, since the similarity X2 is greater than or equal to the threshold value Xs (X2 ≥ Xs; YES in S7), the control unit 10 displays, on the display unit 6 for operator approval, the reference table pattern 71 in a new table format together with the table pattern 61 extracted by the extraction unit 2 (S8).

[0064] The reference table pattern 71 in the new table format displayed on the display unit 6 has a different external shape from the table pattern 61 displayed on the display unit 6. Since the reference table pattern 71 in the new table format is different from the table pattern 61, the operator cannot approve the reference table pattern 71 in the new table format as information necessary for the reading process of the form reading device.

[0065] When information indicating that the reference table pattern 71 in the new table format displayed on the display unit 6 is not approved is input by the operator's operation of the operation unit 5 (NO in S9), the control unit 10 notifies the information indicating non-approval by display on the display unit 6 or print output by the print unit 4 (S11). Then, the control unit 10 proceeds to the reading process of the reading unit 1 for the next form (S1).

[0066] When the detected similarity X2 is less than the threshold value Xs (X2 < Xs; NO in S7), the control unit 10 notifies the information to that effect by display on the display unit 6 or print output by the print unit 4 (S11). Then, the control unit 10 proceeds to the reading process of the reading unit 1 for the next form (S1).

[0067] As described above, it is possible to grasp the table formats of various forms set in the reading unit 1. Therefore, the contents of the tables on various forms can be read quickly and efficiently. Moreover, since the determination of whether to register is left to the judgment of the operator, the reliability regarding the registration of the table format can be enhanced.

[0068] The present invention is not limited to the above-described embodiments, and at the implementation stage, the components can be modified and embodied without departing from the gist thereof. Also, various inventions can be formed by appropriately combining a plurality of components disclosed in the above-described embodiments.

Explanation of Reference Numerals

[0069] 1...reading unit, 2...extraction unit, 3...storage unit, 4...printing unit, 5...operation unit, 6...display unit, 10...control unit.

Claims

1. a reading unit that reads a form including a table; a storage unit that stores a reference table pattern in a predetermined table format; a control unit that recognizes the table of the document read by the reading unit and the information in the table based on a reference table pattern stored in the storage unit; Equipped with The control unit a first detection means for detecting a first similarity between the table of the form read by the reading unit and a reference table pattern stored in the storage unit; a generating means for generating a new table-format reference table pattern that includes, as a table, a portion of the reference table pattern that corresponds to the table of the form read by the reading unit, but does not include other portions as tables, when the first similarity detected by the detecting means is less than a threshold value; a second detecting means for detecting a second similarity between the table of the form read by the reading unit and the new table-format reference table pattern generated by the generating means; a control means for storing a new tabular reference table pattern generated by the generating means in the storage unit when the second similarity detected by the second detecting means is equal to or greater than the threshold value; Including, Document reader.

2. the generating means generates the new reference table pattern in a table format, when the first similarity detected by the detecting means is less than the threshold value, under the judgment that the table format of the form read by the reading unit is different from the table format of the reference table pattern stored in the storage unit. The document reading device according to claim 1.

3. The control unit a control means for determining, when the first similarity detected by the first detection means is equal to or greater than the threshold value, that the table format of the form read by the reading unit is the same as the table format of the reference table pattern in the storage unit, and proceeding to a reading process of the reading unit for the next form; further comprising: The document reader according to claim 2 .

4. The device further includes a display unit, an operation unit, and a printing unit, The control means If the second similarity detected by the second detecting means is equal to or greater than the threshold value, a new reference table pattern in tabular form generated by the generating means is displayed on the display unit for operator approval; when an operator inputs information approving the displayed new tabular reference table pattern by operating the operation unit, the new tabular reference table pattern generated by the generation means is stored in the storage unit, and the reading unit proceeds to a reading process for the next document; when the operator inputs information indicating that the displayed new tabular reference table pattern is not approved by operating the operation unit, the information indicating that the operator does not approve is displayed on the display unit or printed out by the printing unit, and the reading unit proceeds to read processing for the next document; If the second similarity detected by the second detection means is less than the threshold value, information to that effect is displayed on the display unit or printed out by the printing unit, and the reading unit proceeds to a reading process for the next document. The document reading device according to claim 1.

5. a reading unit that reads a form including a table; a storage unit that stores a reference table pattern in a predetermined table format; a control unit that recognizes the table of the document read by the reading unit and the information in the table based on a reference table pattern stored in the storage unit; A control method for a document reading device comprising: Detecting a first similarity between the table of the form read by the reading unit and a reference table pattern in the storage unit; If the detected first similarity is less than a threshold value, a new table-format reference table pattern is generated, which includes, as a table, a portion of the reference table pattern corresponding to the table of the form read by the reading unit, but does not include other portions as tables; Detecting a second similarity between the table of the form read by the reading unit and the generated new reference table pattern in table format; If the detected second similarity is equal to or greater than the threshold, the generated new reference table pattern in tabular form is stored in the storage unit. A method for controlling a document reading device.

6. a reading unit that reads a form including a table; a storage unit that stores a reference table pattern in a predetermined table format; a control unit configured by a computer, which recognizes the table of the form read by the reading unit and the information in the table based on the reference table pattern; A program for a document reading device comprising: The computer, a function of detecting a first similarity between the table of the form read by the reading unit and a reference table pattern in the storage unit; a function of generating a new table-format reference table pattern that includes, as a table, a portion of the reference table pattern that corresponds to the table of the form read by the reading unit and does not include other portions as tables, when the detected first similarity is less than a threshold value; a function of detecting a second similarity between the table of the form read by the reading unit and the newly generated reference table pattern in table format; a function of storing the generated new tabular reference table pattern in the storage unit when the detected second similarity is equal to or greater than the threshold value; A program for a document reading device that achieves this.

Images