JP5223739B2 - Portable character recognition device, character recognition program, and character recognition method - Google Patents

Description

  The present invention relates to a portable character recognition apparatus, a character recognition program, and a character recognition method that recognize characters included in image data, for example.

Conventionally, there are applications using various character recognition such as business card reading and translation readers.
In general, when using these applications, a user acquires a target character to be recognized as an image and recognizes the character from the acquired image.

  In order to perform character recognition more accurately, the user sets the target language in advance and adjusts the camera shooting mode such as whether to use the light and macro lens to acquire images and the degree of zooming to recognize characters. Make settings to obtain images suitable for

  By the way, in general, a user often uses the same application for a roughly determined purpose according to a place (position), and therefore, it is considered that a user often makes similar settings for a target language and a camera shooting mode. .

  For example, when a Japanese user goes to China on a business trip or trip and checks what is written on the menu at a restaurant, the target language is set to Chinese, for example, the light is on and a macro lens is used The camera shooting mode is set, and the zoom degree is adjusted so that the size of the character string to be recognized is within a predetermined range required by the application.

  Repeatedly go to the same place (location) for business trips and trips, for example, “I was delicious when I went last time”, the same restaurant, the user needs to make the same settings when using the same application again .

  Also, in public facilities such as airports, when checking what is written on the information board, the user sets the target language to Chinese, for example, the camera shooting mode that does not turn on the light and does not use the macro mode Then, the zoom degree is adjusted so that the size of the character string to be recognized is within a predetermined range required by the application. Similarly, when the user repeatedly goes to the same place (position) and uses the same application again, the user needs to make the same setting.

  As described above, in general, it is considered that the user often sets the target language and the camera shooting mode roughly according to the location (position), but the user intends to use the application. Each time, it was necessary to make the same setting with the same troublesome work.

  In such a situation, there has been proposed an apparatus that acquires a current position using a GPS device and performs settings related to character recognition based on position information (see Patent Documents 1 and 2).

  The portable terminal of Patent Literature 1 is a device that acquires a current position using a GPS device and sets the character type of a character to be read from the current position in character recognition for character input to an application. In this device, since the character type is automatically set based on the current position acquired by the GPS device, the operation by the user can be reduced.

  The translation device of Patent Document 2 is a device that acquires a current position using a GPS device and prioritizes character types used at the current position in character recognition. This device can improve the recognition accuracy because the character type specified based on the current position acquired by the GPS device is easily recognized preferentially.

  These apparatuses perform settings on software when character recognition processing is performed from an acquired image, and can effectively perform character recognition on an image that has been clearly acquired.

  However, these devices do not reduce the trouble associated with camera shooting modes such as the presence / absence of the use of lights and macro lenses and the degree of zoom, so-called hardware settings, for obtaining a clear image.

  In addition, the setting on the hardware greatly affects the recognition accuracy. For this reason, for example, the above-mentioned device was recognized well before because it was not set even though it was set to use a macro lens when used in the same place (position) before. There was a problem that things could not be recognized.

  Thus, the character recognition process is greatly affected by the accuracy of the acquired image. For this reason, settings on hardware such as a camera shooting mode that greatly affects an acquired image are important. In addition, the above-described apparatus requires troublesome setting work for the user.

JP 2005-141520 A JP 2006-302091 A

  SUMMARY OF THE INVENTION In view of the above-described problems, the present invention provides a portable character recognition device, a character recognition program, and a character recognition method that realize accurate character recognition with a simple operation, and aims to improve user satisfaction. It is said.

The present invention obtains matching data storage means for storing matching data for character recognition, character recognition means for acquiring image data and recognizing characters in the image data based on the matching data, and position information. A portable character recognition device comprising position information acquisition means, wherein the character recognition means sets image information obtained when an image is acquired and character recognition, and the position of the character recognized by the character recognition means Related information storage means for storing position information in association with each other, storage position confirmation means for confirming whether or not the position for character recognition by the character recognition means is stored in the related information storage means, and stored in the related information storage means When the character recognizing means recognizes a character at a certain position, the setting condition information associated with the position information is read out to recognize the character And a condition setting means for setting the stage, accepting a change operation to change from the acquired image data to the character correct character recognition in which was false character erroneously in recognized character recognized by the recognition means, changing the change character Recognizing character changing means, misidentified character information, and changed character information that has been changed, and misidentified character change information storage means for storing the character information in association with the position information as misidentified character change information. Misidentified character confirmation means for confirming whether the recognized character recognized by the character recognition means at the position stored in the related information storage means is stored as misidentified character information in the misidentified character change information storage means; It has been confirmed by the character confirmation means that the recognized character matches the misidentified character stored in the misidentified character change information storage means. The case, characterized by comprising a character correction means for correcting the change characters on the basis of the mistaken character modification information.

The matching data may be constituted by data used for character recognition matching, such as feature data indicating character characteristics.
The storage means can be constituted by means for storing data, such as a hard disk or a nonvolatile memory.

The image data may be appropriate image data such as a color image, a gray scale image, or a monochrome two-tone image.
The position information acquisition means for acquiring the position information can be a GPS device, a device that receives wireless communication from a fixed point, and specifies a position.

  The character recognition means is, for example, an image pickup means such as a CCD camera or a CMOS camera, or a USB (Universal Serial Bus) or LAN (Local Area Network) that receives image data taken by such an image pickup means from the image pickup means. The feature data is extracted from the character image acquired through the communication interface such as), and the feature data is matched with the feature data of the matching data to recognize which character.

  The setting condition information is a setting condition that is set when the character recognition unit acquires an image and recognizes a character, and relates to a setting condition on software such as a character type of a character to be recognized, and an imaging unit for acquiring an image. It can be configured with camera shooting modes such as whether or not a light or macro lens is used and the zoom level, so-called hardware setting conditions.

  The portable character recognition device is constituted by an appropriate device such as a cellular phone provided with an imaging means, a control means, and a storage means, or an information processing terminal such as a PDA (Personal Digital Assistants) or a notebook personal computer. can do.

The recognized character is a character recognized from a character image in the image data captured by the character recognition means, and may be a predetermined appropriate character such as hiragana, katakana, kanji, or a plurality of these.

The misidentified character is a character misrecognized among the recognized characters recognized from the character image in the image data captured by the character recognizing means, and the changed character is a character obtained by correctly changing the misidentified character.

  According to the present invention, accurate character recognition can be realized with a simple operation on image data, and user satisfaction can be improved.

  Specifically, since the setting condition information set at the time of character recognition is stored in association with the position information acquired by the position information acquisition means, the same setting condition is set again when character recognition is performed again at the same location (position). Can be set by reading. Therefore, character recognition can be performed under the same conditions as the previous time without setting the same conditions.

  Note that the same location (position) determined based on the location information acquired by the location information acquisition means can be determined to be the same location (position) within a predetermined range. May be set arbitrarily.

Further , when the same character as the misidentified character is recognized at a location (position) where the misidentified character has been changed to the changed character before, the recognized character can be corrected to the changed character. Therefore, when the same character as the misidentified character is recognized again at the same place (position), it can be corrected to the correct character after the change without performing a change operation, and the convenience for the user is improved.

  Further, as an aspect of the present invention, the character correction unit has a matching reliability with respect to the misidentified character of the recognized character recognized by the character recognition unit within a first predetermined range, and a matching reliability with respect to the changed character. In the second predetermined range, the recognized character recognized by the character recognition means can be corrected to the changed character.

Thereby, when the same character as a misidentified character is recognized, the accuracy of the correction for correcting the character correctly is improved, and accurate character recognition can be executed.
Specifically, at a place (position) where a misidentified character has been changed to a changed character previously, a character whose matching reliability for the misidentified character is within the first predetermined range, which is a character that can be determined to be approximate to the misidentified character, is recognized. In the case where a recognized character is a character that can be determined to be approximate to the changed character, and a character whose matching reliability for the changed character is within the second predetermined range is recognized, the recognized character is corrected to the changed character. .

  Therefore, it is possible to correct a recognized character that approximates to some extent both a misidentified character and a modified character, as well as a recognized character that has been previously misidentified but changed correctly, and only approximates either the misidentified character or the modified character. Compared with the case where the recognized character to be corrected is corrected, the recognized character can be accurately corrected to the changed character, and the character recognition accuracy can be improved.

  Further, as an aspect of the present invention, a first matching reliability of the recognized character recognized by the character recognition unit when the recognized character is changed to the changed character by the recognized character changing unit, and the changed character Matching reliability storage means for storing a second matching reliability for the first matching range, the first predetermined range is set to a predetermined range centered on the first matching reliability, and the second predetermined range is The second matching reliability can be set within a predetermined range.

As a result, when the same character as the misidentified character is recognized, the accuracy of correction to correct the changed character correctly can be further improved, and more accurate character recognition can be performed.
Specifically, since a predetermined range can be set based on the matching reliability between the previously recognized recognized character and the misidentified character, and the matching reliability between the recognized character and the changed character, the misidentification that was once misidentified but correctly changed Correction of the recognized character based on the character change record can be realized.

As an aspect of the present invention, the position acquisition means can be configured by a GPS device.
Thereby, accurate position information can be acquired over a wide range. Therefore, for example, even when characters are recognized overseas, it is possible to accurately recognize the position and recognize the characters, thereby further improving the convenience for the user.

Further, the present invention provides a matching data storage means for storing matching data for character recognition, a character recognition means for acquiring image data and recognizing characters in the image data based on the matching data, and position information. A character recognition program to be executed by a portable terminal comprising position information acquisition means for acquiring, wherein the character recognition means recognizes characters in the image data based on the matching data, and image acquisition is performed. The related information storage step for storing the setting condition information set when the character is recognized and the position information of the position recognized by the character recognition step in the related information storage means, and the character recognition step. A storage position confirmation step for confirming whether a character recognition position is stored in the related information storage means. And flop, when character recognition with the related information stored in the storage means position, to execute a condition setting step of setting the same condition by reading the setting condition information associated with the positional information to the mobile terminal, A change operation accepting step for accepting a change operation for changing a misrecognized character that was erroneously recognized in the recognized character recognized by the character recognizing step from the acquired image data to a correct character, and changing the misrecognized character based on the accepted change operation Recognized character change information storage step of storing in the recognized character change information storage means in association with the position information as misidentified character change information as the recognized character change step, the recognized character change step to change to characters, and the changed character information changed To the mobile terminal, and at the position confirmed in the storage position confirmation step The step of confirming whether or not the recognized character is stored as the misidentified character information in the misidentified character change information storage means, and the recognized character is stored in the misidentified character change information storage means by the misidentified character confirmation step. When it is confirmed that the character matches the misidentified character, the portable terminal is caused to execute a character correction step for correcting the character to be changed based on the misidentified character change information .

  The character recognition program can be stored in an appropriate storage medium such as a memory, a chip, or a hard disk.

  According to the present invention, if a character recognition program is installed in a portable terminal, accurate character recognition can be realized with a simple operation on image data.

  Specifically, since the setting condition information set at the time of character recognition is stored in association with the position information acquired by the position information acquisition means, the mobile terminal installed with the character recognition program is again installed at the same location (position). When recognizing characters, the same setting conditions can be read and set. Therefore, character recognition can be performed under the same conditions as the previous time without setting the same conditions.

In addition , when the same character as the misidentified character is re-recognized by the mobile terminal at the location (position) where the misidentified character is changed to the changed character before, the recognized character can be corrected to the changed character. Therefore, when the same character as the misidentified character is recognized again at the same place (position), it can be corrected to the correct character after the change without performing a change operation, and the convenience for the user is improved.

The present invention also provides a character recognition step for acquiring image data and recognizing characters in the image data based on matching data for character recognition stored in the matching data storage means, and a position for acquiring position information. A character recognition method for executing an information acquisition step, wherein the character recognition step includes setting condition information set when an image is acquired and the character is recognized, and the position information of the position where the character is recognized by the character recognition step. Associated information storage step, a storage position confirmation step for confirming whether or not a character recognition position is stored in the character recognition step, and a character recognition at the stored position. wherein the running and condition setting step, and acquisition by reading the setting condition information which is set to the same conditions A change operation accepting step for accepting a change operation for changing a misrecognized character, which was erroneously recognized in the recognized character recognized by the character recognition step from the image data, to a correct character, and changing the misidentified character to a changed character based on the accepted change operation. A recognized character changing step, and a misidentified character change information storage step of storing the misidentified character information and the changed changed character information as misidentified character change information in association with the position information in the misidentified character change information storage means. And a step of confirming whether the recognized character recognized at the position confirmed in the storage position confirmation step is stored as misidentified character information, and the recognized character stored in the misidentified character confirmation step. When it is confirmed that it matches the misidentified character, it is based on the misidentified character change information. And executes the character correction step of correcting the change characters Te.

  Thus, accurate character recognition can be performed on the image data with a simple operation.

  Specifically, since the setting condition information set at the time of character recognition is stored in association with the position information, the same setting condition can be read and set again at the same place (position) when the character is recognized. Therefore, character recognition can be performed under the same conditions as the previous time without setting the same conditions.

Further , when the same character as the misidentified character is recognized at a location (position) where the misidentified character has been changed to the changed character before, the recognized character can be corrected to the changed character. Therefore, when the same character as the misidentified character is recognized again at the same place (position), it can be corrected to the correct character after the change without performing a change operation, and the convenience for the user is improved.

  According to the present invention, a portable character recognition device, a character recognition program, and a character recognition method that realize accurate character recognition with a simple operation can be provided, and the satisfaction of the user can be improved.

The perspective view of a portable terminal. The block diagram which shows the structure of a portable terminal. Explanatory drawing explaining the structure of dictionary data and related information storage data. The flowchart of the character recognition process by a character recognition program. Explanatory drawing of the setting screen about character recognition and a recognition character.

  An embodiment of the present invention will be described below with reference to the drawings.

  FIG. 1 is a perspective view of the mobile terminal 1, FIG. 2 is a block diagram of the configuration of the mobile terminal 1, and FIG. 3 is an explanatory diagram for explaining the configuration of dictionary data and related information storage data.

  As shown in FIGS. 1 and 2, the mobile terminal 1 includes an image input unit 3, an operation unit 5, a screen display unit 7, and a GPS unit 9. The image input unit 3 is a digital camera that captures an image, the operation unit 5 is a button that is pressed, and the screen display unit 7 is a liquid crystal display or an organic EL display that displays an image.

  As shown in FIG. 2, the mobile terminal 1 includes a CPU 11, a ROM 12, and a RAM 13 connected to a bus 16 in addition to an image input unit 3, an operation unit 5, a screen display unit 7, and a GPS unit 9. Yes.

The image input unit 3 is a digital camera such as a CCD camera or a CMOS camera, and is connected to the bus 16. A captured image obtained by the image input unit 3 is transmitted to the CPU 11 as digital image data. Note that the image input unit 3 receives an imaging signal from the CPU 11 and starts imaging.
The operation unit 5 is connected to the bus 16. The operation unit 5 transmits a pressing signal pressed by the user to the CPU 11 via the bus 16.

The screen display unit 7 is connected to the bus 16, and when image data is sent from the CPU 11, an image of the image data is displayed as a language setting screen 7a, a camera shooting mode setting screen 7b, or a misidentified character change screen 7c. indicate.
The GPS unit 9 can receive radio signals from three or more GPS satellites (not shown), and can calculate the position of the mobile terminal 1 based on the received radio signals.

  The CPU 11 is a processor that performs various operations using the RAM 13 as a temporary storage area in accordance with a program stored in the ROM 12, and has a control function 21 that controls each unit and a character recognition function 22.

  The character recognition function 22 includes a character string extraction processing unit 31, a character extraction processing unit 32, a character recognition processing unit 33, a recognition reliability determination processing unit 34, and a recognition result correction processing unit 35.

  The character string extraction processing unit 31 executes processing for extracting a character string from image data. The extraction of the character string is executed by an appropriate method such as binarizing the image data to obtain projection data and distinguishing the character string and the margin from the projection data. Specifically, a vertical histogram is calculated from the number of black pixels for each pixel row for a character image, and a range in which a predetermined number or more of black pixel rows are continuous is used as a character string. Run.

  The character extraction processing unit 32 executes processing for extracting characters from the character string. For example, in this character extraction, a horizontal histogram is calculated from the number of black pixels for each pixel column for horizontally written characters, and a range in which a predetermined number or more of black pixels are continuous is defined as one character or written vertically A vertical histogram is calculated from the number of black pixels for each pixel row for a character, and a range in which a predetermined number or more of black pixels are continuous is defined as one character.

  The character recognition processing unit 33 performs processing for recognizing characters by determining which character matching information matches or approximates using the matching information of the matching dictionary data 24 and processing for obtaining matching reliability.

The recognition reliability determination processing unit 34 executes processing for determining whether or not the matching reliability (recognition reliability) obtained by the character recognition processing unit 33 exceeds a predetermined range.
The recognition result correction processing unit 35 executes a process of correcting an erroneously recognized character to a changed character that has been correctly changed.

  The ROM 12 performs character recognition processing such as a character recognition program 23, matching dictionary data 24, a predetermined allowable value α that defines a predetermined range for determining matching reliability, which will be described later, and area information for determining the same position. Necessary information is stored.

  As shown in FIG. 3A, the matching dictionary data 24 includes a character code and matching information, and matching dictionary data 24a and 24b exist for each language.

The character code is a character code of an appropriate method such as Shift-JIS, for example.
The matching information stores character feature data for each character.

  The RAM 13 temporarily stores data used by the CPU 11 for various controls and calculations, and stores an image such as a character string image in the image memory 41. Furthermore, as shown in FIG. 3B, related information storage data 42 related to the result of changing the misidentified character to the correct changed character is stored.

  Specifically, as shown in the change information management table 42a in FIG. 3 (B), the change processing is performed in the related information storage data 42, which is information related to the change record obtained by changing a misperformed character to a correct changed character. It is composed of the position information of the position where the image is performed, the target language, the camera shooting mode, and the changed character information.

The position information is position information acquired by the GPS unit 9 and can be position information based on longitude and latitude or coordinate values, for example.
The target language is a language set by the user when characters are recognized, as will be described later.

  The camera shooting mode is a setting of the image input unit 3 when the image input unit 3 captures a misidentified character that has been changed to a correct change character. Whether the light is used, whether the macro lens is used, and how the zoom is used. It consists of the presence or absence of.

  The changed character information includes the misidentified character code, the matching reliability (hereinafter referred to as the first matching reliability) for the misidentified character of the target character portion of the character image captured by the image input unit 3, and the misidentified character as the user. And a matching reliability (hereinafter referred to as a second matching reliability) for the changed character of the target character portion in the image captured by the image input unit 3.

  Since the RAM 13 is composed of a non-volatile memory, the related information storage data 42 can be stored without disappearing even when the power of the mobile terminal 1 is turned off.

  Note that the mobile terminal 1 is also provided with a communication antenna for performing wireless communication, a speaker for outputting sound, a microphone for collecting sound, and the like. Thereby, it is comprised so that it can utilize as a mobile telephone.

  Further, as indicated by virtual lines in FIG. 2, the CPU 11 and the ROM 12 may be mounted on one chip 15 and configured as electronic components. In this case, if the chip 15 is provided in the portable terminal 1, a character recognition function can be easily added.

Next, the character recognition in the portable terminal 1 is demonstrated with FIG.4 and FIG.5.
4 is a flowchart of character recognition processing executed by the CPU 11 of the mobile terminal 1 in accordance with the character recognition program 23. FIG. 5 is an explanatory diagram of a setting screen for recognizing characters and recognized characters.

  In the following explanation, for example, consider a case where a translation application that reads Chinese and translates into Japanese is used to check what is written on the menu at a restaurant while traveling to China. Try.

  More specifically, for example, an attempt was made to examine the meaning of an item “Kaori venison” in a Chinese restaurant written in a slightly different font character as shown in the misidentified character change screen 7c in FIG. A character recognition process in the translation application will be described.

First, when the translation application of the portable terminal 1 is started by the user, the CPU 11 acquires current position information by the GPS unit 9 (step S1). And CPU11 confirms whether the positional information on the acquired present position is the position memorize | stored in the relevant information storage data 42 (step S2).
Note that whether the position information of the current position and the position stored in the related information storage data 42 are the same place (position) is stored in the related information storage data 42 based on the area information stored in the ROM 12. In this configuration, it is determined whether or not the position information of the current position is included in a predetermined range including the determined position.

  It is stored in the related information storage data 42 that is a place (position) that is visited for the first time with the portable terminal 1 or that has been visited before but has no record of changing a character that was mistakenly recognized by the character recognition process in the translation application. When the position is not (step S2: No), the CPU 11 displays a language setting screen 7a for allowing and selecting a recognizable language on the screen display unit 7 (see FIG. 5A).

  The language setting screen 7a includes a selectable language list display unit 71 having a check box 71a for selecting one of a plurality of languages for each language, a return button 72 for returning to the previous screen, and a confirmation button 73. , And a submenu button 74 for displaying a submenu.

  When the user inputs a check in the check box 71a and presses the confirm button 73 on the language setting screen 7a, the CPU 11 matches the matching dictionary data of the corresponding language among the matching dictionary data 24 stored in the ROM 12. 24 is read and the target language is set (step S3).

  As described above, in this embodiment in which the Chinese menu is translated into Japanese, as shown in FIG. 5A, when a check is input to the Chinese check box 71a and the confirm button 73 is pressed. The CPU 11 reads the matching dictionary data (Chinese version) 24b from the ROM 12, and completes the language setting.

  When the language setting is completed, the CPU 11 displays a camera shooting mode setting screen 7b for setting a shooting mode for clearly shooting an image including characters to be authenticated by the image input unit 3 (camera). b) see).

  The camera shooting mode setting screen 7b includes a light setting unit 75 for setting use of a light for shooting an image, a macro lens setting unit 76 for setting use of a macro lens, a zoom setting unit 77 for setting zoom, A return button 72, a confirm button 73, and a submenu button 74 are provided.

The light setting unit 75 includes check boxes 75a for setting whether or not to use, and similarly, the macro lens setting unit 76 includes check boxes 76a for the presence or absence.
The zoom setting unit 77 includes a slide bar 77a that slides in the left-right direction on the screen display unit 7 to adjust the zoom amount between telephoto (T) and wide-angle (W).

  In the camera shooting mode setting screen 7b, when the user inputs a check in the check boxes 75a and 76a, adjusts the slide bar 77a to an appropriate position, and presses the confirm button 73, the CPU 11 is selected. The camera shooting mode is set (step S4).

  By setting in this way, a clear character image is photographed by the image input unit 3 and character recognition of the target language is possible. Therefore, the user who is directed to the character portion to be subjected to character authentication of the portable terminal 1 When the shutter button (not shown) is pressed by the CPU 11, the CPU 11 performs image shooting by the image input unit 3 and acquires a character image (step S 5).

  CPU11 which acquired the character image by the image input part 3 performs the character string extraction process by the character string extraction process part 31, and while extracting a character string from a character image (step S6), about the extracted extracted character string, Character recognition processing is executed by the character recognition processing unit 33 (step S7), and the character recognition processing result is displayed on the screen display unit 7 (see FIG. 5C, step S8).

In this character recognition processing, character recognition is executed using the matching dictionary data 24 for each character extracted from the character string by the character extraction processing unit 32.
Note that character extraction by the character extraction processing unit 32 corresponds to processing for extracting a boundary between characters. For example, the character string image is binarized so that the character portion is black, black pixels are projected in the vertical direction on the character string, and the projected data break is set as a candidate for the boundary between the character and the character.

  In this character recognition processing, for each character extracted by the character recognition processing unit 33, a character having high matching reliability with the target character in the matching dictionary data 24 is calculated as the authentication character 82a. The matching reliability at this time is the first matching reliability described above.

The misrecognized character change screen 7c displaying the result of the character recognition process includes a character display unit 80, a return button 72, a confirmation button 73, and a submenu button 74.
The character display unit 80 displays a character image display unit 81 that displays a character image 81a, a recognized character string display unit 82 that displays a recognized character string 82a that is a recognized character string, and each recognized character 82b of the recognized character string 82a. A change character selection window 83 that displays the change candidate characters 83a in a selectable manner is provided.

  When the recognized character string 82a composed of the same recognized characters 82b as the character image 81a displayed on the character image display unit 81 is displayed on the recognized character string display unit 82 on the misidentified character change screen 7c, all characters are displayed. Since it is correctly recognized (step S9: No), when the user presses the confirmation button 73, the CPU 11 recognizes the recognized character 82b “Kaori venison” recognized from the character image 81a “Kaori venison”. The character recognition process is terminated, and the translation process for the recognized character string 82a is executed.

  On the other hand, if there is a character recognized in error (hereinafter referred to as “misidentified character 82c”) in the character image 81a, the change candidate character 83a of the misidentified character 82c is displayed in the modified character selection window 83 (step S9). : Yes).

  Specifically, as shown in FIG. 5 (c), when a recognition character string 82a "Kokuka deer" is displayed for the character image 81a "Koroku deer", a recognition character "Koroshi deer" is displayed. In the column 82a, the recognized characters 82b of “incense”, “mochi”, and “deer” are correctly recognized, and “meat” is erroneously recognized as the misidentified character 82c “month”.

As described above, when there is a misidentified character 82c in the recognized character string 82a, when the user selects the "month" that is the misidentified character 82c, the CPU 11 changes the change candidate character 83a of "month" that is the misidentified character 82c. It is displayed in the character selection window 83.
The change candidate character 83a displayed in the changed character selection window 83 displays a character with high matching reliability calculated in step S7 described above.

  Then, when “meat” which is the correct change character 83b is selected from the change candidate characters 83a and the confirm button 73 is pressed, the CPU 11 recognizes “Kaori venison” recognized from the character image 81a “Kaori venison”. "Is recognized (step S10).

  CPU11 reads the matching reliability with respect to the change character 83b of the misrecognized part 81b calculated as step S7 and misidentified as the misrecognized character 82c in the character image 81a, and makes it 2nd matching reliability as mentioned above (step S11). ). The CPU 11 may calculate the matching reliability between the misidentified portion 81b and the changed character 83b in the character image 81a.

  Further, the CPU 11 registers the information obtained by changing the recognized character 82b to the changed character 83b in the related information storage data 42 (step S12), ends this character recognition processing, and subsequently executes the translation processing for the recognized recognized character string 82a. To do.

  As shown in FIG. 4B, the contents to be registered in step S12 include the position information acquired in step S1, the target language set in step S3, the camera shooting mode setting contents set in step S4, and The changed character information and the like are stored in the changed information management table 42a.

  The changed character information stored in the changed information management table 42a includes the character code of the erroneously recognized character 82c and the first matching reliability between the misidentified portion 81b and the erroneously recognized character 82c in the character image 81a calculated in step S7. The character code of the changed character 83b changed in step S10 and the second matching reliability for the changed character 83b of the misidentified portion 81b read out in step S11 are registered.

  On the other hand, when the position information acquired in step S1 is the position stored in the related information storage data 42 (step S2: Yes), the CPU 11 reads and sets the setting language stored in the related information storage data 42 ( At the same time as step S13), the setting details of the camera shooting mode stored in the related information storage data 42 are set (step S14).

  In this state, the image is captured by the image input unit 3 under the same condition as the previous character recognition setting condition stored in the related information storage data 42, so that the character can be recognized. Similar to S5 to S7, a character image is acquired by the image input unit 3, a character string extraction process by the character string extraction processing unit 31, and a character recognition process by the character recognition processing unit 33 is executed (step S15). Note that the matching reliability with the target character in the character recognition process is referred to as “third matching reliability”.

Then, as a result of the character recognition process, the CPU 11 determines whether there is a recognized character 82b stored in the related information storage data 42 as a misidentified character in the recognized character string 82a (step S16).
When the CPU 11 determines that the recognized character 82b as a result of the character recognition process does not include the misidentified character 82c stored in the related information storage data 42 as the misidentified character (step S16: No), the process proceeds to step S8.

  On the other hand, when it is determined that the recognized character 82b as a result of the character recognition process includes the misidentified character 82c stored in the related information storage data 42 as the misidentified character (step S16: Yes), the CPU 11 calculates in step S15. The recognized reliability determination processing unit 34 determines whether the third matching reliability is within a first predetermined range having a predetermined allowable value α1 with the first matching reliability calculated in step S7 as a central value. When it is determined that the matching reliability is not within the first predetermined range (step S17: No), the process proceeds to step S8.

When the third matching reliability is within the first predetermined range (step S17: Yes), the CPU 11 determines that the misrecognized character 82c is stored as the misrecognized character in the related information storage data 42, and the related information storage. The matching reliability with the changed character 83b stored in the data 42 is calculated (step S18).
The matching reliability at this time is referred to as “fourth matching reliability”.

  Then, the CPU 11 determines whether the fourth matching reliability is within the second predetermined range having the predetermined allowable value α2 with the second matching reliability calculated in step S11 as a central value by the recognition reliability determination processing unit 34. If it is determined that the fourth matching reliability is not within the second predetermined range (step S19: No), the process proceeds to step S8.

  When the fourth matching reliability is within the second predetermined range (step S19: Yes), the CPU 11 corrects the recognition result 82c for the misidentified character 82c determined to be stored as the misidentified character in the related information storage data 42. The processing unit 35 corrects the changed character 83b stored in the related information storage data 42 (step S20).

  Furthermore, the CPU 11 determines whether or not there is a recognized character stored as a misidentified character in the related information storage data 42 in the remaining recognized character 82b in the recognized character string 82a, and if it does not remain (step S21: No), step S8. If it remains (step S22: Yes), the process returns to step S17.

  As described above, the CPU 11 of the mobile terminal 1 can accurately recognize the character from the character image acquired by the image input unit 3 based on the above-described flowchart, and can execute the translation process on the recognized recognized character string 82a. .

  In addition, by using the portable terminal 1 that performs the character recognition process, it is possible to acquire a clear image with a simple operation and to realize accurate character recognition. Further, when the same recognized character 82b as the misidentified character 82c is recognized at the position (position) where the misidentified character 82c has been changed to the changed character 83b before, the character is corrected to the changed character 83b without operating the changed character selection window 83 or the like. This improves user convenience.

  Specifically, if you repeatedly go to the same place (location) on a business trip, you may go to the same restaurant repeatedly, for example, because it was delicious the last time you went, or repeatedly use the same public facilities Conceivable.

  In such a case, as shown in the flowchart above, when the target language or camera shooting mode set when the character recognition process is executed for the first time at a certain place (position) goes to the same place (position) again and the character is recognized. Since it is automatically set in conjunction with the GPS unit 9 for acquiring position information, it is not necessary to repeat troublesome labor.

  As a result, the macro lens was set to be used when it was previously used at the same location (position), such as when traveling on a business trip or traveling abroad, but it was not set. Can no longer recognize what was well recognized.

In addition, regarding the misrecognized character 82c changed to the changed character 83b, when a character is recognized from the same character image 81a at the same place (position), there is a very high possibility that it is erroneously recognized. However, since the mobile terminal 1 that performs the character recognition process is automatically changed based on the related information stored in the related information storage data 42 in conjunction with the GPS unit 9 without performing the same changing operation again. , User convenience is improved.
Note that the possibility of misrecognition when recognizing the same character at the same location (position) is more likely in the case of a special font that is not supported.

  For example, as shown in the character image 81a in the above-described embodiment (see FIG. 5C), the left side of the character “meat” written with characters of a slightly different font is similar to the left side of the character “month”. For this reason, it was mistakenly recognized as “Moon” and had to be changed to the correct “Meat”.

  Note that the character fonts in the menu are generally uniform, and in this restaurant, the same misrecognition occurs when the character “meat” is recognized. However, since the misrecognized character 82c and the changed character 83b are stored in the related information storage data 42 for each place (position), the character can be recognized as the correct changed character 83b without a troublesome change operation.

As a measure for improving the recognition performance in the character recognition process, it is conceivable to use a character that could not be recognized to cause the matching dictionary to learn.
According to this method, in general, the character shape stored in the matching dictionary by learning is between the corresponding font and the character font that could not be recognized. If there is a large difference in the font shape of a character that has not been received, there is a risk that the recognition accuracy for a character that has had good recognition accuracy will deteriorate. That is, by learning, when an application is used in another place (position), it is conceivable that characters that had good recognition accuracy when they came last time cannot be recognized much.

  On the other hand, in the above-described character authentication process, the third matching reliability calculated in step S15 is based on the first matching reliability of the misidentified character 82c stored in the related information storage data 42 together with the position information as a central value. The second matching reliability calculated in step S18 and the second matching reliability for the changed character 83b stored in the related information storage data 42 together with the position information is a second value centered on the first predetermined range. In the predetermined range, it is possible to accurately correct the misidentified character 82c to the changed character 83b based on the results of the change processing previously performed at the same place (position).

  Therefore, when the last time it came, characters that had good recognition accuracy would not be recognized so much, and there was no inconvenience that the erroneously recognized characters 82c that were once mistakenly changed to the changed characters 83b were changed. Based on this, it is possible to perform accurate character recognition.

  In the above embodiment, when a character mistakenly recognized by the character recognition process by the mobile terminal 1 is changed, the position information, the target language, the setting content of the camera shooting mode, and the changed character information are stored in the related information storage data 42. However, the position information, the target language, and the camera shooting mode may be stored in the related information storage data 42 each time the character recognition process is executed.

In this case, after setting the target language in step S3 and setting the camera shooting mode in step S4, the target language and the camera shooting mode are stored in the related information storage data 42 together with the position information acquired in step S1.
Thereby, for example, even when the character misidentified in the character recognition processing by the mobile terminal 1 is not changed, the same setting is made at the same location (position) as the location information stored in the related information storage data 42. Can do.

  In addition, when comprised in this way, the information amount memorize | stored in the relevant information storage data 42 increases, and there exists a possibility that extraction of the change character information for correction | amendment of the misidentified character 82c may become slow. Therefore, the management table for managing the position information, the target language, and the camera shooting mode stored each time the character recognition process is executed, and the management table different from the position information and the changed character information as the results when the misidentified character 82c is changed. It is preferable to manage.

In correspondence between the configuration of the present invention and the above-described embodiment,
The matching data of the present invention corresponds to the matching dictionary data 24,
Similarly,
The image data corresponds to the character image 81a,
The matching data storage means corresponds to the ROM 12,
The character recognition means and the character recognition step correspond to steps S5 to S7 and S15,
The position information acquisition means and the GPS device correspond to the GPS unit 9,
The portable character recognition device and the portable terminal correspond to the portable terminal 1,
The setting condition information corresponds to the target language and camera shooting mode stored in the change information management table 42a.
The related information storage means and the related information storage step correspond to step S12,
The storage position confirmation means and the storage position confirmation step correspond to step S2,
The condition setting means and the condition setting step correspond to steps S13 and S14,
The recognized character changing means, the change operation accepting step and the recognized character changing step correspond to step S10,
The misidentified character change information corresponds to the changed character information stored in the change information management table 42a,
The misidentified character change information storage means and the misidentified character change information storage step correspond to step S12,
The erroneous character confirmation means and the erroneous character confirmation step correspond to step S16,
The character correction means and the character correction step correspond to step S20.
The present invention is not limited only to the configuration of the above-described embodiment, and many embodiments can be obtained.

1 ... mobile terminal 9 ... GPS unit 12 ... ROM
23 ... Character recognition program 24 ... Matching dictionary data 24b ... Matching dictionary data (Chinese version)
42a ... changed information management table 81a ... character image 82c ... misidentified character 82b ... recognized character 83b ... changed character

Claims (6)

Matching data storage means for storing matching data for character recognition;
Character recognition means for acquiring image data and recognizing characters in the image data based on the matching data;
A portable character recognition device comprising position information acquisition means for acquiring position information,
Related information storage means for storing the setting condition information set when the character recognition means obtains an image and character recognition and the position information of the position recognized by the character recognition means in association with each other;
A storage position confirmation means for confirming whether or not the position for character recognition by the character recognition means is stored in the related information storage means;
A condition setting unit that reads the setting condition information associated with the position information and sets the character recognition unit when the character recognition unit recognizes a character at the position stored in the related information storage unit ;
A recognition character changing unit that accepts a change operation to change a misrecognized character that was erroneously recognized in the recognized character recognized by the character recognizing unit from the acquired image data to a correct character;
A misidentified character change information storage means for storing the misidentified character information and the changed modified character information as misidentified character change information in association with the position information;
A misidentification that confirms whether the recognized character recognized by the character recognition unit at the position stored in the related information storage unit and stored in the related information storage unit is stored as misidentified character information in the misidentified character change information storage unit. Character verification means;
Character correcting means for correcting to the changed character based on the misidentified character change information when the recognized character confirmation means confirms that the recognized character matches the misidentified character stored in the misidentified character change information storage means. A portable character recognition device. The character correcting means;
The character recognition means when the matching reliability of the recognized character recognized by the character recognition means with respect to the misidentified character is within a first predetermined range and the matching reliability with respect to the changed character is within a second predetermined range. The portable character recognition device according to claim 1, wherein the recognized character recognized by the character is corrected to the changed character. A first matching reliability for the misrecognized character of the recognized character recognized by the character recognizing unit when the misrecognized character is changed to the changed character by the recognized character changing unit, and a second matching reliability for the changed character. A matching reliability storage means for storing,
Setting the first predetermined range to a predetermined range centered on the first matching reliability;
Mobile character recognition apparatus according to the second predetermined range, to claim 2 which is set to a predetermined range around the second matching reliability. The portable character recognition device according to any one of claims 1 to 3 , wherein the position acquisition means is configured by a GPS device. Matching data storage means for storing matching data for character recognition;
Character recognition means for obtaining image data and recognizing characters in the image data based on the matching data;
A character recognition program to be executed by a mobile terminal provided with position information acquisition means for acquiring position information,
In the character recognition step, wherein the character recognition means recognizes characters in the image data based on the matching data.
A related information storage step of associating and storing in the related information storage means the setting condition information set at the time of image acquisition and character recognition, and the position information of the position recognized by the character recognition step;
A storage position confirmation step for confirming whether or not a position for character recognition by the character recognition step is stored in the related information storage means;
When recognizing characters at the position stored in the related information storage unit, the portable terminal is caused to execute a condition setting step of reading the setting condition information associated with the position information and setting the same condition .
A change operation accepting step for accepting a change operation for changing a misrecognized character that was erroneously recognized in the recognized character recognized by the character recognizing step from the acquired image data to a correct character;
A recognized character changing step for changing a misidentified character to a changed character based on the received change operation;
While causing the portable terminal to execute a misidentified character change information storage step of storing the misidentified character information and the changed modified character information as misidentified character change information in association with the position information and storing the misidentified character change information in the storage unit,
A misidentified character confirmation step for confirming whether the recognized character recognized at the position confirmed in the storage position confirmation step is stored as misidentified character information in the misidentified character change information storage means;
A character correction step for correcting to the changed character based on the misidentified character change information when it is confirmed that the recognized character matches the misidentified character stored in the misidentified character change information storage means in the misidentified character confirmation step. A character recognition program that runs on a mobile device . A character recognition step of acquiring image data and recognizing characters in the image data based on matching data for character recognition stored in the matching data storage means;
A character recognition method for executing a position information acquisition step of acquiring position information,
In the character recognition step,
A related information storage step for storing the setting condition information set when the image is acquired and the character is recognized and the position information of the position where the character is recognized in the character recognition step in association with each other;
A storage position confirmation step for confirming whether a character recognition position is stored in the character recognition step;
A condition setting step of reading the setting condition information associated with the position information and setting the same condition when recognizing characters at the stored position ;
A change operation accepting step for accepting a change operation for changing a misrecognized character that was erroneously recognized in the recognized character recognized by the character recognizing step from the acquired image data to a correct character;
A recognized character changing step for changing a misidentified character to a changed character based on the received change operation;
Performing the misidentified character change information storage step of storing the misidentified character information and the changed modified character information as misidentified character change information in association with the position information and storing in the misidentified character change information storage means;
A misidentified character confirmation step for confirming whether the recognized character recognized at the position confirmed in the storage position confirmation step is stored as misidentified character information;
A character correction step for correcting the changed character to the changed character based on the misidentified character change information when the recognized character is confirmed to match the stored recognized character by the step of checking the recognized character. Recognition method.

Images