JP6222541B2 - Image processing apparatus and program - Google Patents

Description

  The present invention relates to an image processing apparatus and a program.

  Patent Document 1 discloses a keyword character string described in the vicinity of a title character string on a document image, extraction condition acquisition means for acquiring relative position information of the title character string with respect to the keyword character string, Character recognition means for recognizing at least a region including the title character string and the keyword character string, and a keyword character string acquired by the extraction condition acquisition means is searched from the recognition result of the character recognition means. Based on the keyword search means for acquiring the position, the position of the keyword character string acquired by the keyword search means, and the relative position of the title character string with respect to the keyword character string acquired by the extraction condition acquisition means Title position acquisition means for acquiring the position of the character string, and the title position acquired by the title position acquisition means Based on the position of the character string, discloses title extracting apparatus characterized by having a title output means for outputting the data of the title string.

  Patent Document 2 relates to a plurality of items, the entry fields and preprinted item names are arranged, and the entry fields are obtained from a form image obtained by reading a form in which the entry fields and the item names are partitioned by ruled lines. In a form OCR program that causes a computer to execute a form OCR process for recognizing characters in the form, the OCR process is performed on the entire surface of the form image, and the position of the entry frame that divides the entry field and the items surrounding the item name Full-frame OCR processing step for recognizing the position of name frames and character strings in these frames, and an entry field specifying processing step for specifying an entry field to be subjected to OCR processing again based on the result information of the full-screen OCR processing step And a partial OCR processing step that partially executes OCR processing on the specified entry field. It discloses R program.

JP 2008-077454 A JP 2005-173730 A

  An object of the present invention is to provide an image processing apparatus and program capable of improving the extraction accuracy of a desired character string from a document image obtained by reading a recording medium.

  According to the first aspect of the present invention, an analysis unit that analyzes a document image obtained by reading a recording medium, and detects coordinates of a plurality of predetermined character strings in the document image from an analysis result by the analysis unit. A coordinate detection unit; a plurality of predetermined character string coordinates included in a reference image corresponding to the document image; a target point coordinate included in the reference image; and a plurality of coordinates detected by the coordinate detection unit. Based on the coordinates, the coordinate acquisition means for acquiring the coordinates of the target point of the document image corresponding to the target point of the reference image, and the coordinates acquired by the coordinate acquisition means from the analysis result by the analysis means And a character string extracting unit that extracts a character string in the region.

  The present invention according to claim 2 is the image processing apparatus according to claim 1, wherein the coordinate detecting means detects coordinates of a plurality of predetermined character strings using a regular expression.

  The present invention according to claim 3 is characterized in that the coordinate acquisition means uses the coordinates of the plurality of predetermined character strings included in the reference image and the coordinates detected by the coordinate detection means. The image processing apparatus according to claim 1, wherein a magnification of enlargement or reduction between the document image and the document image is acquired, and coordinates of a target point of the document image are acquired based on the magnification.

  According to a fourth aspect of the present invention, the coordinate acquisition unit is configured to change the coordinates of the plurality of predetermined character strings included in the reference image to a target point included in the reference image. A plurality of coordinates of the target point of the document image corresponding to the target point of the reference image are acquired, and the character string extraction unit is any one of a plurality of regions based on the coordinates acquired by the coordinate acquisition unit. The image processing apparatus according to claim 1, wherein a character string included in the area is extracted.

  The present invention according to claim 5 is the image processing apparatus according to claim 4, wherein the character string extraction means extracts character strings included in all of a plurality of regions based on the coordinates acquired by the coordinate acquisition means. It is.

  The present invention according to claim 6 is the image processing apparatus according to claim 4 or 5, wherein the character string extracting means extracts a character string that satisfies a predetermined condition.

  In the present invention according to claim 7, the coordinate detection unit detects coordinates of at least three predetermined character strings, and the coordinate acquisition unit detects the at least three predetermined characters included in the reference image. The coordinate of the target point of the document image is acquired based on the coordinates of the character string and the coordinates of the target point in the region surrounded by the coordinates and the plurality of coordinates detected by the coordinate detecting means. 2. The image processing apparatus according to 2.

  According to an eighth aspect of the present invention, an analysis step for analyzing a document image obtained by reading a recording medium, and coordinates of a plurality of predetermined character strings in the document image are detected from the analysis result in the analysis step. The coordinate detection step, the coordinates of the plurality of predetermined character strings included in the reference image corresponding to the document image, the coordinates of the target point included in the reference image, and the plurality detected in the coordinate detection step. The coordinate acquisition step of acquiring the coordinates of the target point of the document image corresponding to the target point of the reference image based on the coordinates of the reference image, and the coordinates acquired in the coordinate acquisition step from the analysis result in the analysis step Is a program that causes a computer to execute a character string extraction step of extracting a character string in a region based on.

  According to the first aspect of the present invention, there is provided an image processing apparatus capable of improving the extraction accuracy of a desired character string from a document image obtained by reading a recording medium as compared with a case where the present configuration is not provided. can do.

  According to the second aspect of the present invention, in addition to the effect of the present invention according to the first aspect, an image capable of efficiently detecting a predetermined character string coordinate used for extracting a desired character string. A processing device can be provided.

  According to the present invention of claim 3, in addition to the effect of the present invention of claim 1 or 2, even if the document image is enlarged or reduced compared to the original image, a desired character string Can be provided.

  According to the present invention of claim 4, in addition to the effect of the present invention of claim 1 or 2, the positional relationship between the character strings included in the document image is broken from the positional relationship in the original image. Even if it exists, the image processing apparatus which can extract a desired character string can be provided.

  According to the fifth aspect of the present invention, in addition to the effect of the present invention according to the fourth aspect, an image processing apparatus that can extract a desired character string more reliably than the case without the present configuration is provided. can do.

  According to the sixth aspect of the present invention, in addition to the effect of the present invention according to the fourth or fifth aspect, an image processing apparatus that can extract a desired character string more reliably than the case without the present configuration. Can be provided.

  According to the present invention of claim 7, in addition to the effect of the present invention of claim 1 or 2, a desired character string can be obtained even when the document image is enlarged or reduced compared to the original image. Can be provided.

  According to the eighth aspect of the present invention, there is provided a program capable of improving the extraction accuracy of a desired character string from a document image obtained by reading a recording medium as compared with a case where the present configuration is not provided. Can do.

It is a schematic diagram which shows the hardware constitutions of the image processing apparatus 2 which concerns on embodiment of this invention. It is a block diagram which shows the function structure of the image processing apparatus 2 implement | achieved by running a program. It is a schematic diagram explaining extraction of the character string of this embodiment, (a) illustrates a reference image, (b) illustrates information stored in the reference information storage unit 34 in advance, (c ) Exemplifies a document image. 4 is a flowchart showing a flow of storing reference information in a reference information storage unit 34; It is a flowchart which illustrates about operation | movement which extracts a character string from a document image. It is a schematic diagram explaining extraction of a character string in the first modification, (a) exemplifies a reference image, (b) exemplifies information stored in advance in the reference information storage unit 34, (C) illustrates a document image. It is a flowchart which illustrates about operation | movement which extracts a character string from a document image in a 1st modification. It is a schematic diagram explaining extraction of a character string in the second modification, (a) illustrates a reference image, (b) is an area surrounded by coordinate points of three use character strings in the reference image The positional relationship between the target points is illustrated. FIG. 10 is a schematic diagram for explaining extraction of a character string in a second modification, where (a) illustrates a document image, and (b) is an area surrounded by coordinate points of three use character strings in the document image. The positional relationship between the target points is illustrated. 14 is a flowchart illustrating an operation of extracting a character string from a document image in the second modification.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a schematic diagram showing a hardware configuration of an image processing apparatus 2 according to an embodiment of the present invention.

  As shown in FIG. 1, the image processing apparatus 2 includes a main body 8 including a CPU 4 and a memory 6, an input device 10, an output device 12, a storage device 14, a storage medium 16, a communication device 18, and an image reading device 20. The image processing apparatus 2 includes a component as a computer that can communicate with other apparatuses.

  The CPU 4 executes processing based on the program stored in the memory 6. The storage device 14 is an internal HDD or the like, and the storage medium 16 is a CD, FD, external HDD, or the like. Note that the CPU 4 may execute a program stored in the storage device 14 or the storage medium 16.

  The input device 10 is a keyboard, a mouse, and a joystick, and the output device 12 is a display device such as a display. Note that the input device 10 and the output device 12 may be configured by a touch panel or the like.

  The communication device 18 is a communication device such as a data line termination device that communicates with other devices via a communication line such as a LAN or the Internet.

  The image reading device 20 is an image reading device such as a scanner, and reads a recording medium such as paper as a document image.

  FIG. 2 is a block diagram illustrating a functional configuration of the image processing apparatus 2 realized by executing the program. Note that part or all of the configuration shown in FIG. 2 may be realized by hardware such as ASIC and FPGA.

  As shown in FIG. 2, the image processing apparatus 2 includes a document image acquisition unit 30, an analysis unit 32, a reference information storage unit 34, a coordinate detection unit 36, a coordinate acquisition unit 38, and a character string extraction unit 40. The character string is extracted from the document image obtained by reading the recording medium.

  The document image acquisition unit 30 acquires document image data obtained by reading a recording medium. In the present embodiment, the document image acquisition unit 30 acquires the document image data read by the image reading device 20 from the image reading device 20, but the document image acquisition unit 30 is transmitted via, for example, the communication device 18. The document image data may be acquired, or the document image data stored in the storage medium 16 may be read out and acquired. In the following description, the document image indicates document image data acquired by the document image acquisition unit 30.

  The analysis unit 32 analyzes the document image data acquired by the document image acquisition unit 30. The analysis unit 32 analyzes the document image by using, for example, an OCR (Optical Character Recognition) function, and detects characters and character positions included in the document image.

  The reference information storage unit 34 stores information about the reference image corresponding to the document image acquired by the document image acquisition unit 30. The reference image is a predetermined image corresponding to the document image acquired by the document image acquisition unit 30. For example, image data that is a formation target when the document image is formed on a recording medium, and a template of the document This corresponds to image data. Details of the information stored in the reference information storage unit 34 will be described later.

  Here, the position of the character string included in the reference image in the reference image may be different from the position of the character string in the document image. For example, when the image reading device 20 reads the recording medium with a deviation, the position of the character string in the document image is deviated from the position in the reference image. Even when the image is enlarged or reduced compared to the reference image, the position of the character string in the document image is different from the position in the reference image. In such a case, even if an attempt is made to extract the character string from the document image acquired by the document image acquisition unit 30 using absolute coordinates in the reference image of the character string (for example, coordinates with the upper left corner of the image as the origin). Since the absolute coordinates in the document image of the character string have changed, there is a possibility that they cannot be extracted.

  In the present embodiment, when a desired character string is extracted from the document image acquired by the document image acquisition unit 30, a character string to be extracted is not designated by absolute coordinates, but from a plurality of predetermined character strings. Specify the character string to be extracted using relative coordinates. In the following description, this predetermined character string may be referred to as a use character string.

  The coordinate detection unit 36 detects the coordinates (absolute coordinates) of the plurality of predetermined character strings (used character strings) in the document image acquired by the document image acquisition unit 30 from the analysis result of the analysis unit 32. Note that what the character string is used is stored in the reference information storage unit 34 in advance.

  The coordinate acquisition unit 38 includes the coordinates (absolute coordinates) of the plurality of predetermined character strings included in the reference image corresponding to the document image acquired by the document image acquisition unit 30 and the target points included in the reference image. Based on the coordinates and the plurality of coordinates detected by the coordinate detection unit 36, the coordinates of the target point of the document image corresponding to the target point of the reference image are acquired. The target point is a point indicating the position of the character string to be extracted. In this embodiment, the coordinates of the target point included in the reference image will be described as relative coordinates from the coordinates of a predetermined character string. However, the relative coordinates can be calculated if the absolute coordinates are known. It may be a coordinate.

  Note that the coordinates of the used character string included in the reference image corresponding to the document image acquired by the document image acquisition unit 30 and the coordinates of the target point included in the reference image are stored in the reference information storage unit 34 in advance.

  The character string extraction unit 40 extracts a character string in a region based on the coordinates of the target point acquired by the coordinate acquisition unit 38 from the analysis result of the analysis unit 32. In the present embodiment, a character string in a rectangular area using the coordinates of the target point as a base point is extracted. Note that the width and height of the rectangular region are determined based on the information about the width and height stored in advance in the reference information storage unit 34.

Next, extraction of a character string in the present embodiment will be described using a specific example.
FIG. 3 is a schematic diagram for explaining extraction of a character string according to the present embodiment. FIG. 3A illustrates a reference image, and FIG. 3B is stored in the reference information storage unit 34 in advance. FIG. 3C illustrates a document image. Here, as a result of facsimile transmission of the recording medium on which the reference image is formed, the document image as the facsimile reception image is reduced as compared with the reference image, and the entire character string position is moved in parallel. An example will be described.

  In the example shown in FIG. 3, it is assumed that the character string to be extracted is “789123” described as the order number, and a plurality of predetermined strings used for extraction of this character string are used. It is assumed that there are two character strings (utilization character strings) of “Order Form” and “Everest”.

  In FIG. 3A, the + mark indicates the coordinate position of the used character string, and the frame line indicates the area of the character string to be extracted in the reference image. In this example, the upper left corner point of the frame line is the target point in the reference image, but the target point may be the other corner point of the frame line.

  As shown in FIG. 3B, for example, the reference information storage unit 34 uses the coordinates of the used character string in the reference image, the coordinates of the target point in the reference image, and the width and height of the rectangular area based on the target point. Information is stored.

  In the example shown in FIG. 3, the coordinates in the reference image of the usage character string “order” are (x1_org, y1_org), and the coordinates in the reference image of the usage character string “Everest” are (x2_org, y2_org). The relative coordinates of the target point in the reference image (relative coordinates from the coordinates of the use character string “Order Form”) are (x_org, y_org), the width of the frame line in the reference image is w_org, Assume that the line height is h_org.

  Similarly, the coordinates in the document image of the used character string “Purchase Order” are (x1_img, y1_img), the coordinates in the document image of the used character string “Everest” are (x2_img, y2_img), and the document image The relative coordinates of the target point in (the relative coordinates from the coordinates of the usage string “Purchase Order”) are (x_img, y_img), the width of the border in the document image is w_img, and the height of the border is , H_img.

  In FIG. 3C, the + mark is a projection of the coordinate position of the used character string in the reference image, and the x mark indicates the coordinate position of the used character string in the document image. Also, the broken frame line is a projection of the frame line in the reference image, and the solid line frame line indicates the area of the character string to be extracted in the document image. As described above, as a result of the reduction and movement of the document image as compared with the reference image, the character string “789123” to be extracted does not exist within the dashed frame.

  In the example illustrated in FIG. 3, the coordinate detection unit 36 detects the coordinates (x1_img, y1_img) and (x2_img, y2_img) of the used character string from the OCR result of the analysis unit 32 for the document image.

  The coordinate acquisition unit 38 acquires the relative coordinates (x_img, y_img) of the target point in the document image, for example, by the following calculation formula.

  x_img = (x_org + (x1_img-x1_org)) * (| x2_img-x1_img | / | x2org-x1_org |) y_img = (y_org + (y1_img-y1_org)) * (| y2_img-y1_img | / | y2org-y1_org |)

  As described above, the coordinate acquisition unit 38 uses the coordinates of a plurality of predetermined character strings (utilized character strings) included in the reference image and the coordinates detected by the coordinate detection unit 36 to use the reference image and the document image. The magnification of enlargement or reduction between and is acquired, and the coordinates of the target point of the document image are obtained based on this magnification.

  On the other hand, the character string extraction unit 40 extracts a character string included in an area based on (x_img, y_img) in the document image from the OCR result of the analysis unit 32. Note that the width and height of the region with (x_img, y_img) as the base point are w_img and h_img. Here, the values of w_img and h_img are calculated, for example, by multiplying the width w_org and the height h_org of the frame line in the reference image by the magnification of the document image with respect to the reference image.

  As described above, in the example shown in FIG. 3, “order form” and “everest” are used as the character strings used, but the character strings used are, for example, in advance in the document image among the character strings included in the reference image. It is preferable that the character string is expected to exist in a predetermined area. For example, it is preferable that a character string or the like previously included in the document template data such as a document title or an item name is selected as the use character string.

  In the example shown in FIG. 3, an example of using a unique character string such as “Order Form” or “Everest” as the use character string is shown. However, a character string represented by a regular expression may be used as the use character string. Good. For example, by using a meta character, a character string corresponding to a character string composed of three kanji characters, a character string having an arbitrary character after the character string of “order”, or the like may be designated as a use character string. Good. For example, when the coordinates of the used character string “order book” are detected by the coordinate detection unit 36, the character of the “book” cannot be recognized, and the character string “order book” is searched from the analysis result of the analysis unit 32. Even if it is not possible, if you specify a character string with an arbitrary character after the character string of "order" by regular expression instead of "order form", you can find the usage character string Become.

FIG. 4 is a flowchart showing a flow of storing reference information in the reference information storage unit 34.
In step 10 (S10), a use character string is determined. For example, the user designates which of the character strings included in the reference image is to be used as the use character string, and the image processing apparatus 2 determines the designated character string as the use character string. In the present embodiment, a plurality of character strings are designated as the use character strings.

  In step 12 (S12), relative coordinates to the character string to be extracted from the used character string determined in step 10 are designated. For example, relative coordinates are determined by designating a target point by the user. Further, the width and height of the extraction area with the target point as the base point are designated by the user.

  In step 14 (S14), the coordinate information for each character string used, the relative coordinates of the target point, and the width and height information of the extraction area based on the target point are stored in the reference information storage unit 34.

FIG. 5 is a flowchart illustrating an operation for extracting a character string from a document image.
In step 20 (S20), the document image acquisition unit 30 acquires the document image read by the image reading device 20.

  In step 22 (S22), the analysis unit 32 analyzes characters and character positions included in the document image.

  In step 24 (S24), the reference information stored in the reference information storage unit 34 is read.

  In step 26 (S26), it is determined whether or not the used character string defined in the reference information exists in the document image. If the used character string exists, the process proceeds to step 28. Goes to step 34.

  In step 28 (S28), the coordinate detector 36 detects the absolute coordinates of the used character string in the document image.

  In step 30 (S30), the coordinate acquisition unit 38 uses the coordinates of the used character string and the target point in the reference image defined in the reference information, and the coordinates in the document image of the used character string detected in step 28. The relative coordinates of the target point in the document image are acquired.

  In step 32 (S32), the character string extraction unit 40 extracts the character string in the region based on the target point from the analysis result in step 22 based on the relative coordinates acquired in step 30.

  On the other hand, in step 34 (S34), since character strings cannot be extracted based on relative coordinates, character strings in an area based on predetermined absolute coordinates are extracted.

  In the embodiment described above, a plurality of used character strings are provided. However, the number of used character strings is one, and the character string extraction region in the document image is determined by the relative distance from the one used character string to the target point. You may make it ask. However, if there is one character string used, it may not be possible when the document image is enlarged or reduced relative to the reference image.

  Next, a first modification of the present embodiment will be described. In the above embodiment, the reference information storage unit 34 stores relative coordinates from any of the plurality of used character strings as the coordinates of the target point. On the other hand, in the first modification, relative coordinates from each of a plurality of used character strings are stored. In the first modification, the coordinate acquisition unit 38 corresponds to the target point of the reference image according to the relative coordinates from the coordinates of the plurality of used character strings included in the reference image to the target point included in the reference image. A plurality of coordinates of the target point of the document image to be acquired, and the character string extraction unit 40 selects a character string included in any one of the plurality of regions based on the coordinates acquired by the coordinate acquisition unit 38. Extract.

  FIG. 6 is a schematic diagram for explaining extraction of a character string in the first modification. FIG. 6A illustrates a reference image, and FIG. 6B is stored in the reference information storage unit 34 in advance. FIG. 6C illustrates a document image.

  Here, a case will be described as an example where the arrangement of character strings in the document image differs from the arrangement of character strings in the reference image, for example, by converting the data format of the document data. For example, when the reference image is an image based on PDF (Portable Document Format) document data, the document data is converted into other format document data, and the converted document data is imaged on a recording medium. In this case, the arrangement of the character strings of the document image read from the recording medium may be different from the arrangement of the character strings in the reference image. In the example shown in FIG. 6, in the document image, the character string of “Order Form” designated as the use character string is shifted to the left as compared with the reference image. As a result, the positional relationship between the character string of “Order Form” in the reference image and the character string below “Company Name”, and the positional relationship between the character string of “Order Form” in the document image and the character string below “Company Name” Is different. In the example shown in FIG. 6, the character string in the document image is further shifted in the lower right direction as a whole.

  In the example shown in FIG. 6, the document image is not enlarged or reduced with respect to the reference image. However, the process shown as the first modification may be applied to the enlarged or reduced document image.

  In the example shown in FIG. 6, it is assumed that the character string to be extracted is “789123” described as the order number, and a plurality of predetermined characters used for extraction of this character string. It is assumed that there are three columns (utilization character strings): “Order Form”, “Everest”, and “Page”.

  In FIG. 6A, the + mark indicates the coordinate position of the used character string, and the frame line indicates the area of the character string to be extracted in the reference image. In this example, the point at the upper left corner of the frame line is the target point in the reference image. In addition, each arrow extending from the + mark to the target point clearly indicates a vector that clearly indicates the relative coordinates from each coordinate position of the used character string to the target point.

  As shown in FIG. 6B, for example, the reference information storage unit 34 uses the coordinates of the used character string in the reference image, the coordinates of the target point in the reference image, and the width and height of the rectangular area based on the target point. Information is stored. Here, the first modification is different from the above-described embodiment in that the relative coordinates from each of the plurality of used character strings are stored.

  In FIG. 6C, the x mark indicates the coordinate position of the used character string in the document image. Each arrow extending from the x mark is a projection of the arrow shown in FIG. 6A (a vector indicating relative coordinates from the coordinate position of the used character string to the target point), and is a vector represented by the arrow. Is equal to the vector shown in FIG. 6A and the vector shown in FIG. In FIG. 6C, the end point of the vector indicates the coordinate position of the target point of the document image corresponding to the target point of the reference image, and the frame line with the target point as the base point in the document image is shown in FIG. It is the same as the frame line shown in (a). As described above, in the reference image shown in FIG. 6A and the document image shown in FIG. 6C, the vectors and the frame lines are the same, but the coordinate position of the used character string is changed. Therefore, the areas surrounded by the frame lines are different.

  In the first modification, the coordinate acquisition unit 38 uses the coordinates of the used character string in the document image detected by the coordinate detection unit 36 and the relative coordinates to the target point stored in the reference information storage unit 34. The coordinates of the target point of the document image are obtained from the used character string. In the example illustrated in FIG. 6, the coordinate acquisition unit 38 uses the coordinates of the used character string “order form” in the document image detected by the coordinate detecting unit 36 and the used character string “stored in the reference information storage unit 34. The coordinates of the target point from the use character string “order sheet” are acquired based on the coordinates in the reference image of the “ordering document” and the relative coordinates from the use character string to the target point. Similarly, the coordinate acquisition unit 38 also acquires the coordinates of the target point from the usage character string “Everest” and the coordinates of the target point from the usage character string “page”.

  In the first modification, the character string extraction unit 40 extracts a character string included in any one of a plurality of regions based on the coordinates acquired by the coordinate acquisition unit 38. In the example shown in FIG. 6, first, an area is specified based on the coordinates of the used character string “order book” acquired by the coordinate acquisition unit 38. The designation of the area is performed by determining the rectangular area using the width and height information of the rectangular area stored in the reference information storage unit 34. Similarly, an area based on the coordinates of other used character strings acquired by the coordinate acquisition unit 38 is also specified. Each designated region is represented as a region within each frame line in FIG. Next, the character string extraction unit 40 extracts, for example, character strings included in all designated areas as extraction target character strings.

  Note that the character string extraction unit 40 may extract character strings included in all areas specified in this way as extraction target character strings, but are included in a number of areas equal to or greater than a predetermined threshold. A character string may be extracted as a character string to be extracted.

  The character string extraction unit 40 may extract a character string that satisfies a predetermined condition. For example, it may be possible to extract only a character string having a predetermined character type such as a number or alphabet, or analyze the contents of the character string in the region by performing morphological analysis or the like. Alternatively, only a character string having a predetermined content (for example, a character string indicating an address or a character string indicating a name) may be extracted.

  FIG. 7 is a flowchart illustrating an operation of extracting a character string from a document image in the first modification. Here, since Step 20 to Step 24 are the same as those in the flowchart in FIG. 4, Step 30 and subsequent steps will be described.

  In step 30 (S30), it is determined whether or not the use character string defined in the reference information exists in the document image. If the use character string exists, the process proceeds to step 32. Goes to step 38.

  In step 32 (S32), the coordinate detection unit 36 detects the absolute coordinates of the used character string in the document image.

  In step 34 (S34), the coordinate acquisition unit 38 calculates the document image from the coordinates of the used character string in the document image detected in step 32 and the relative coordinates to the target point stored in the reference information storage unit 34. The relative coordinates of the target point of the target character string are obtained.

  In step 36 (S 36), the character string extraction unit 40 extracts the character string in the region based on the coordinates acquired in step 34 from the analysis result of step 22.

  In step 38 (S38), it is determined whether or not another use character string is defined in the reference information. If another use character string is defined, the process returns to step 30 and the other use character string is defined. The above processing is performed on the character string. On the other hand, if the above processing has been performed for all defined use character strings, the process proceeds to step 40.

  In step 40 (S40), it is determined whether or not there is a use character string determined to be present in the document image in step 30, and it is determined in step 30 that one or more use character strings exist in the document image. If there is no character string to be used that is determined to exist in the document image, the process proceeds to step 44.

  In step 42 (S42), the character string extraction unit 40 determines a character string from the character strings extracted in step 36. For example, as described above, the character string extraction unit 40 determines character strings included in all designated areas as extraction target character strings. If none of the character strings extracted in step 36 satisfies the condition (for example, it is included only in a part of the designated area), the process may proceed to step 44.

  On the other hand, in step 44 (S44), since character strings cannot be extracted based on relative coordinates, character strings in an area based on predetermined absolute coordinates are extracted.

  In the flowchart described above, in step 40, it is described that if there is one or more character strings determined in step 30 as being present in the document image, the process proceeds to step 42. However, the number is not limited to one. It is good also as 2 or more.

  Next, a second modification of the present embodiment will be described. In the second modification, the coordinate detection unit 36 detects the coordinates of at least three usage character strings, and the coordinate acquisition unit 38 stores at least three usages included in the reference image stored in the reference information storage unit 34. The coordinates of the target point of the document image are acquired based on the coordinates of the character string, the coordinates of the target point in the area surrounded by the coordinates, and the coordinates of the plurality of used character strings detected by the coordinate detection unit 36.

  FIG. 8 is a schematic diagram for explaining extraction of a character string in the second modified example, FIG. 8A illustrates a reference image, and FIG. 8B illustrates three use character strings in the reference image. The positional relationship of the target points within the area surrounded by the coordinate points is shown.

  In FIG. 8A, the + mark indicates the coordinate position of the used character string, the circle mark indicates the target point, and the frame line is an area having the target point as a base point, and the character string to be extracted in the reference image Shows the area. Thus, in the second embodiment, at least three use character strings are used so as to surround the target point.

  In the second modification, the reference information storage unit 34 includes the coordinates of at least three used character strings in the reference image, the coordinate information of the target point in the reference image, and the width and height of the rectangular area based on the target point. Information is stored. For example, as the coordinate information of the target point in the reference image, for example, as shown in FIG. 8B, coordinates indicating the relative position of the target point with respect to the triangle formed by the coordinate points of the use character string are stored. . In the example shown in FIG. 8B, the position of the internal dividing point of the one side by the perpendicular from the target point toward any one side of the triangle (the distance a from the intersection of the perpendicular to the apex shown in the figure) And b), and the position of the internal dividing point of the other side by a straight line toward the other side so as to pass through the target point from one vertex of the one side (shown in the figure from the intersection of the straight line to the vertex) The reference information storage unit 34 stores the distances c and d) as the coordinate information of the target point.

  FIG. 9 is a schematic diagram for explaining extraction of a character string in the second modified example, FIG. 9A illustrates a document image, and FIG. 9B illustrates three used character strings in the document image. The positional relationship of the target points within the area surrounded by the coordinate points is shown. Here, an example will be described in which the document image is reduced as compared with the reference image and the position of the character string is moved in parallel as a whole.

  In FIG. 9A, the x mark indicates the coordinate position of the used character string in the document image, the circle mark indicates the target point in the document image, and the frame line is an area based on the target point. The character string area to be extracted is shown. In the second modification, as shown in FIG. 9 (b), the intersection of the perpendiculars (perpendicular line shown in the figure) so that the internal ratio of the vertical and straight sides is the same as in FIG. 8 (b). The distances a ′ and b ′ from the intersections to the vertices) and the intersections of the straight lines (distances c ′ and d ′ from the intersections of the straight lines to the vertices shown in the figure) are determined, and the position of the target point in the document image is determined. Desired. The coordinate acquisition unit 38 calculates the coordinates of the target point in this way, and the character string extraction unit 40 extracts the character string within the frame line with the target point as the base point, as in the above-described embodiment.

  FIG. 10 is a flowchart illustrating an operation of extracting a character string from a document image in the second modification. Here, since Step 20 to Step 24 are the same as those in the flowchart in FIG. 4, Step 40 and subsequent steps will be described.

  In step 40 (S40), it is determined whether or not the three used character strings defined in the reference information exist in the document image. If there is a used character string, the process proceeds to step 42 and does not exist. If yes, go to step 48.

  In step 42 (S42), the coordinate detector 36 detects the absolute coordinates of each character string used in the document image.

  In step 44 (S44), the coordinate acquisition unit 38 uses the coordinates of the used character string in the document image detected in step 42, the coordinates of the used character string of the reference image stored in the reference information storage unit 34, and the target point. The coordinate of the target point of the document image is acquired from the coordinate information of the document image.

  In step 46 (S 46), the character string extraction unit 40 extracts a character string in the region based on the coordinates acquired in step 44 from the analysis result in step 22.

  On the other hand, in step 48 (S48), a character string in an area based on predetermined absolute coordinates is extracted.

  As described above, in the examples shown in FIGS. 8 and 9, the description has been made using one target point and one triangular area composed of the coordinate points of the used character string. A plurality of different triangular regions surrounding a point may be used. Further, in a plurality of different triangular regions, the target points in each triangular region may be the same vertex on the frame line, or may be different vertices. In this way, when four or more use character strings are set and a plurality of different triangular regions surrounding the target point are used, compared with the case where only three use character strings are set, “Yes” is determined in step 40. It is expected to become easier.

2 Image processing device 4 CPU
6 Memory 20 Image Reading Device 30 Document Image Acquisition Unit 32 Analysis Unit 34 Reference Information Storage Unit 36 Coordinate Detection Unit 38 Coordinate Acquisition Unit 40 Character String Extraction Unit

Claims (7)

An analysis means for analyzing a document image obtained by reading a recording medium;
Coordinate detection means for detecting coordinates of a plurality of predetermined character strings in the document image from the analysis result by the analysis means;
Based on the coordinates of the plurality of predetermined character strings included in the reference image corresponding to the document image, the coordinates of the target point included in the reference image, and the plurality of coordinates detected by the coordinate detection unit. Coordinate acquisition means for acquiring the coordinates of the target point of the document image corresponding to the target point of the reference image;
A character string extracting means for extracting a character string of a region based on the coordinates acquired by the coordinate acquiring means from the analysis result by the analyzing means;
The coordinate acquisition means uses the reference coordinates of the plurality of predetermined character strings included in the reference image and the coordinates of a plurality of character strings corresponding to the document image detected by the coordinate detection means. An image processing apparatus that obtains a magnification for enlargement or reduction between an image and the document image, and obtains coordinates of a target point of the document image based on the magnification. The image processing apparatus according to claim 1, wherein the coordinate detection unit detects coordinates of a plurality of predetermined character strings using a regular expression. The coordinate acquisition means corresponds to the target point of the reference image according to the relative coordinates from the coordinates of the predetermined character strings included in the reference image to the target point included in the reference image. Get multiple coordinates of the target point of the document image,
The image processing apparatus according to claim 1, wherein the character string extraction unit extracts a character string included in any one of a plurality of regions based on the coordinates acquired by the coordinate acquisition unit. The image processing apparatus according to claim 3, wherein the character string extraction unit extracts a character string included in all of a plurality of regions with each coordinate acquired by the coordinate acquisition unit as a base point. The image processing apparatus according to claim 3, wherein the character string extraction unit extracts a character string that satisfies a predetermined condition. The coordinate detecting means detects coordinates of at least three predetermined character strings;
The coordinate acquisition means includes the coordinates of the at least three predetermined character strings included in the reference image, the coordinates of a target point in an area surrounded by the coordinates, and a plurality of coordinates detected by the coordinate detection means. The image processing apparatus according to claim 1, wherein the coordinates of the target point of the document image are acquired based on. An analysis step for analyzing a document image obtained by reading a recording medium;
A coordinate detection step of detecting coordinates of a plurality of predetermined character strings in the document image from the analysis result in the analysis step;
Based on the coordinates of the plurality of predetermined character strings included in the reference image corresponding to the document image, the coordinates of the target point included in the reference image, and the plurality of coordinates detected in the coordinate detection step. A coordinate acquisition step of acquiring the coordinates of the target point of the document image corresponding to the target point of the reference image;
From the analysis result in the analysis step, a program for causing a computer to execute a character string extraction step for extracting a character string in a region based on the coordinates acquired in the coordinate acquisition step,
The coordinate acquisition step uses the reference coordinates of the plurality of predetermined character strings included in the reference image and the coordinates of a plurality of character strings corresponding to the document image detected by the coordinate detection step. A program for obtaining a magnification for enlargement or reduction between an image and the document image and obtaining coordinates of a target point of the document image based on the magnification.

Images