Character recognition device, character recognition method, and recording medium

Abstract

The present invention relates to a character recognition device, a character recognition method, and a recording medium. The character recognition device is provided with: a touch panel type display section (17) that accepts a character image based on an input operation in a handwriting form; and a control section (21) that performs a recognition process on the accepted character image to derive a first character from the character image, displays the derived first character as a candidate for a character desired by a user on the touch panel type display section (17), and in the touch panel type display section (17), with respect to the first character, can accept a designation operation that indicates a meaning of a character similar to the character desired by the user although it is not the character desired by the user, and the control section (21) further derives a second character having a given association with the first character when the designation operation is accepted, and displays the derived second character as an additional candidate for the character desired by the user on the touch panel type display section (17).

Description

Technical Field

[0001] This invention relates to a character recognition device, a character recognition method, and a recording medium. Background Technology

[0002] A technique is proposed that allows for easy correction of a series of handwritten input strings to the desired candidate character without requiring sequential display and selection of each candidate character. (For example, Japanese Patent Application Publication No. 2008-299431)

[0003] Including the technology disclosed in Japanese Patent Application Publication No. 2008-299431, when the desired character is not found among the multiple candidate characters for correction, it is necessary to perform handwritten character recognition processing again. Summary of the Invention

[0004] The present invention is made in view of the above-described facts, and its object is to provide a character recognition device, a character recognition method, and a recording medium that can effectively utilize handwritten input and accept the selection of desired characters from a wide range of candidates.

[0005] One aspect of the present invention is characterized by comprising: an input unit that receives a character image based on a handwritten input operation; a character derivation unit that performs recognition processing on the character image received by the input unit and derives a first character from the character image; and a display control unit that displays the first character derived by the character derivation unit as a candidate character for a user-desired character, wherein the input unit can accept a specified operation regarding the first character that indicates a character similar to the user's desired character, even if it is not the user's desired character; when the specified operation is received by the input unit, the character derivation unit further derives a second character that has a given association with the first character; and the display control unit displays the second character derived by the character derivation unit as an additional candidate character for the user-desired character.

[0006] The effects of the invention

[0007] According to the present invention, the content of handwriting input can be effectively utilized to accept the selection of the desired character from a wide range of candidates. Attached Figure Description

[0008] Figure 1 This is a diagram showing the external structure of an electronic dictionary according to one embodiment of the present invention.

[0009] Figure 2 This is a block diagram illustrating the functional structure of the electronic circuitry of the electronic dictionary involved in this embodiment.

[0010] Figure 3 This is a flowchart illustrating the processing of handwritten input characters according to this embodiment.

[0011] Figure 4 This indicates the implementation method involved. Figure 3 The flowchart shows the detailed subroutine for stroke data correction processing.

[0012] Figure 5 This is a diagram illustrating the handwritten character input and its recognition results according to this embodiment.

[0013] Figure 6 This diagram illustrates the process of exporting multiple corrected stroke data as described in this embodiment.

[0014] Figure 7 This diagram illustrates the transition of the screen on the touch panel display unit according to this embodiment.

[0015] Figure 8 This diagram illustrates the transition of the screen on the touch panel display unit according to this embodiment.

[0016] Figure 9 This is a diagram illustrating a portion of a radical-type table stored in various table storage areas according to this embodiment.

[0017] Figure 10 This diagram illustrates the process of deriving radical-type Chinese character candidates from stroke data input by the user, as described in this embodiment.

[0018] Figure 11 This diagram illustrates an example of the constituent elements involved in this embodiment that replace user-input stroke data with the closest character shape. Detailed Implementation

[0019] The following description, with reference to the accompanying drawings, illustrates one embodiment of applying the present invention to an electronic dictionary.

[0020] [structure]

[0021] Figure 1 This is a front view showing the external structure of the electronic dictionary 10 according to this embodiment. Furthermore, this embodiment can be configured not only as the electronic dictionary 10 described below, but also as a tablet PC (Personal Computer), smartphone, e-book reader, portable game console, server on a communication network, etc., that has dictionary functionality.

[0022] The electronic dictionary 10 is a foldable housing consisting of a main body housing 11 and a cover housing 12 that can be unfolded / closed via a hinge portion 13. On the surface of the main body housing 11 when the foldable housing is unfolded, there are key input sections 14 (hardware keys) including a [main menu] key 14a, a function assignment key 14b, a character input key 14c, a [confirm] key 14d, a [return] key 14e, a [BOX] key 14f, a cursor movement key 14g, and a [switch] key 14h, a sound output section (including a speaker) 15, and a sound input section (including a microphone) 16.

[0023] Furthermore, a touch panel display 17 is provided on the surface of the cover housing 12. The touch panel display 17 is a structure in which a touch position detection device for detecting the position of the user's pen or finger touching the electronic dictionary 10 and a display device are integrated, and it is constructed by overlaying a transparent touch panel on a backlit liquid crystal screen. That is, the touch panel display 17 functions as an input unit for accepting character images based on handwriting input operations.

[0024] The [Main Menu] key 14a of the key input unit 14 is used to display the main menu screen on the key input unit 14. Although not specifically illustrated, it displays multiple icons corresponding to the initial settings or user operations of this electronic dictionary 10. Each icon is a graphic or symbol used to perform the function corresponding to the icon's marking. For example, it includes application icons that directly launch functions (applications) using dictionary content or learning content, and group icons that display a list of icons belonging to multiple functions within a single category.

[0025] The function specification keys 14b of the key input section 14 are keys used to directly specify the dictionary contents, etc., marked on each key. The function specification keys 14b include: a [Multi-Dictionary] key for categories where the dictionary contents are not specifically defined, a key for specifying the category ([Japanese] key, [Ancient Language] key, [Chinese-Japanese] key, [English-English-Japanese] key, etc.), a [Contents Overview] key for displaying a list of dictionary contents, and a [Study Book] key for one category as a tool.

[0026] Furthermore, when the keys on the key input unit 14 are operated after the [Switch] key 14h, they do not function as keys without a frame mark on top, but rather as keys with a frame mark. For example, if the [Delete] key is operated after the [Switch] key 14h, it does not function as the [Delete] key, but rather as the [Set] key.

[0027] exist Figure 1 The text shows the status of selecting the [Japanese dictionary] and choosing the kana input label from the kana input label (あいう) and letter input label (ABC) in the search string input section 41 located at the top of the touch panel display 17.

[0028] The string "[sunlight]" has been entered into the string input section 41. The candidate display section 42 below the string input section 41 shows a list of, for example, three entries corresponding to the string "[sunlight]". Furthermore, the handwriting input section 43 below the candidate display section 42, which displays two characters, is left blank in preparation for the handwriting input of the character following the string "[sunlight]".

[0029] Figure 2 This is a block diagram representing the functional structure of the electronic circuitry of the electronic dictionary 10.

[0030] The electronic dictionary 10's electronic circuitry includes a control unit (CPU: Central Processing Unit) 21, which functions as a computer. The control unit 21 controls the operation of each part of the circuitry according to a control program pre-stored in a memory unit 22, such as a flash ROM. There may be one CPU or more processors.

[0031] In addition, at least one of the control program and the dictionary data described later can be read from an external recording medium 23 such as a memory card by the recording medium reading unit 24 and stored in the storage unit 22, or a portion downloaded from an external network such as a web server (not shown) via the communication unit 25 can be stored in the storage unit 22.

[0032] In addition to being connected to the storage unit 22, the recording medium reading unit 24, and the communication unit 25 via the data bus and the control bus, the control unit 21 is also connected to the key input unit 14, the sound output unit 15, the sound input unit 16, and the touch panel display unit 17.

[0033] In addition to storing the system program responsible for the overall operation of this electronic dictionary 10 and the communication program for communicating with external devices via the communication unit 25, the storage unit 22 also ensures the storage of programs and data used by this electronic dictionary 10 to perform various functions, such as the search processing program storage area 22a, the dictionary data storage area 22b, the search history storage area 22c, and the operation data storage area 22d.

[0034] The retrieval processing program storage area 22a is an area that stores the control program and various table storage areas 22a1, etc., used when performing retrievals corresponding to inputs other than character inputs of entries. The control program retrieves the desired entries and the corresponding translations, meanings, usage examples, explanations, and other explanatory information based on various dictionary contents (English-English-Japanese dictionary / English-Japanese dictionary / Japanese-English dictionary / English-English dictionary / Japanese dictionary / encyclopedia / ...) stored as dictionary data in the dictionary data storage area 22b, and displays them.

[0035] In various table storage areas 22a1, a plurality of tables storing a radical type table including how each character data of Chinese characters can be divided into constituent elements are stored.

[0036] In the retrieval history storage area 22c, when displaying a search term and its explanatory information as a search result in accordance with a search corresponding to a user operation, information on the search term to be searched and its search count is stored together as a retrieval history.

[0037] In the job data storage area 22d, in accordance with the control of the operations of each circuit unit following a control program implemented by the control unit 21, various data such as data input corresponding to a user operation, data obtained or generated by the control unit 21, etc. are temporarily stored (held) as needed.

[0038] The communication unit 25, for example, not only wirelessly connects to external devices such as the same electronic dictionary in the vicinity based on the BLE (Bluetooth (registered trademark) Low Energy) standard, which is one of the wireless PAN (Personal Area Network) technologies, but also wirelessly connects to a Web server, etc. on an external network not shown via a wireless router, etc., to download a control program, etc.

[0039] [Operation]

[0040] Next, the operation of this embodiment will be described.

[0041] In this embodiment, the operation of the correction process in the case where the result of character image recognition of handwritten input is not the Chinese character desired by the user when selecting the function of a Japanese dictionary will be described.

[0042] Here, for example, the following case will be described: The character image of the string [Nikko] is input as a search string by handwritten input and recognized as expected. Then, the character image of the character [Yu] is input by handwritten input in the same way, and the desired character [Yu] is not in the character candidates of the recognition result of the input character image.

[0043] Figure 5 of (A) represents from Figure 1 The screen shows the state where the user of the electronic dictionary 10 further writes the character image of the character [Yu] with the stylus pen P in the handwritten character input unit 43 from the display state of the touch panel type display unit 17 shown.

[0044] As shown in the figure, the handwritten character input unit 43 has an input area for two characters. On the other hand, for example, after performing a pen input by handwriting on the left side, the execution of the recognition process for the input character is indicated using the stylus pen P on the right side.

[0045] In response to an execution instruction for recognition processing using the stylus P, the control unit 21 performs character recognition processing and displays a character candidate list, which is the recognition result, in descending order of evaluated similarity. That is, the control unit 21 functions as a character derivation unit and a display control unit. The character derivation unit performs recognition processing on the character image received in the input unit and derives a first character from the character image. The display control unit displays the first character derived by the character derivation unit as a candidate for the character desired by the user.

[0046] Figure 5 (B) shows a state in which Chinese character candidates with high evaluated similarity in the candidate character display unit 44 are listed. Since the desired Chinese character [浴] is not included in the candidate character display unit 44, the user needs to perform an operation to correct the handwritten input character.

[0047] Figure 3 It is a flowchart showing the processing content of the handwritten input character performed by the control unit 21.

[0048] At the beginning of the processing, the control unit 21 acquires information on candidate characters that match as the recognition result (step S101), and as Figure 5 shown in (B) of, these candidate characters are listed in the candidate character display unit 44.

[0049] Starting from the state in which these Chinese character candidates are listed, the control unit 21 waits for the input of any one of them by repeatedly determining whether a long press operation has been performed on the candidate considered to be the closest (step S103), whether a candidate has been selected (step S107), and whether an operation to cancel the list display has been performed (step S109).

[0050] In step S107, when it is determined that an operation to cancel the list display has been performed (step S109 "Yes"), the control unit 21 ends Figure 3 the processing here.

[0051] Figure 7 (A) of shows a state in which the user has long pressed the Chinese character candidate [洽] in the handwritten character input unit 43 from the screen of the touch panel type display unit 17 shown in Figure 5 (B) of.

[0052] In step S103, when it is determined that a long press operation has been performed on one candidate, the control unit 21 performs a stroke data correction process (step S104) in order to derive multiple corrected stroke data based on the original input stroke data of the handwritten input and the character shape of the candidate Chinese character selected as having the highest similarity.

[0053] Figure 6This is a diagram illustrating the process of deriving multiple corrected stroke data. Figure 6 The stroke data of the handwritten input character shown in (A) of Figure 6 is combined with the shape of the character [Qia] selected by the user as the closest match shown in (B) of Figure 6 Based on this combination, the control unit 21 derives the corrected stroke data 1 shown in (C) of Figure 6 and the corrected stroke data 2 shown in (D) of

[0054] Figure 4 This is a flowchart showing the detailed subroutine of the stroke data correction process executed by the control unit 21 in step S104. At the beginning, the input stroke data of the user is scanned to derive the glyph analysis data including the outer shape frame, blank information, and density information (step S201).

[0055] Based on the derived glyph analysis data, the control unit 21 refers to the radical type table stored in the various table storage areas 22a1 to derive the radical type candidates for the stroke data input by the user (step S202).

[0056] Figure 9 This is a diagram illustrating a part of the radical type table stored in the various table storage areas 22a1. As shown in this diagram, the radical structure of Chinese characters can be classified into Figure 9 the structure composed of "side" and "radical" shown in (A) of Figure 9 the structure composed of "cap" and "foot" shown in (B) of Figure 9 the structure including "wrap" shown in (C) of Figure 9 the structure including "hang" shown in (D) of Figure 9 the structure including other "structures" shown in (E) of Figure 9 other structures not classified as radicals shown in (F) of

[0057] Figure 10 This is a diagram explaining the process of deriving radical type Chinese character candidates from the stroke data input by the user. As Figure 10 shown in (A) of Figure 5 For the stroke data input by the user shown in (A) of Figure 6 scanning is performed sequentially in one direction in the vertical and horizontal directions respectively.

[0058] Figure 10 (B) of Figure 10 shows the results of deriving the outer shape frame OF, blank information SI, and density information CI of the stroke data based on the density of the character image according to the scanning results. Based on these derived results, as Figure 9The result composed of the two radicals, partial Rix1 and side Rix2, shown in (A) considers the Chinese character candidates for long-press operations to derive highly similar radical-type candidates.

[0059] In Figure 10 (C), the stroke data recognized as input by the user is Figure 9 the structure composed of "partial" and "side" shown in (A). On the other hand, in Figure 7 (A), the Chinese character candidate [qia] handwritten by the user with the stylus P in the handwritten character input unit 43 is also Figure 9 the structure composed of "partial" and "side" shown in (A). Thus, when the radical type judged from the handwritten input character image is consistent with the radical type of the character selected as being closer to the glyph, the radical type candidate is limited to one. On the other hand, when the radical type judged from the handwritten input character image is inconsistent with the radical type of the character selected as being closer to the glyph, each radical type is derived as a radical type candidate. That is, the control unit 21 that derives two radical type candidates selects one unselected candidate from the derived radical type candidates (step S203). If the control unit 21 selects one unselected candidate, for example, Figure 9 "partial" and "side" shown in (A), for example, [2] is derived based on [partial] and [side] as the number of constituent elements of the stroke data for the selected radical type candidate (step S204).

[0060] Next, the control unit 21 selects one of the constituent elements input by the user, for example, [partial] (step S205). The control unit 21 selects one of the constituent elements of the character [qia] closest to the one selected by the user for the selected constituent element [partial], for example, [three dots] (step S206).

[0061] The control unit 21 further determines whether the selected constituent element, for example, [three dots], can omit the subsequent replacement and other processes (step S207). This is to reduce the processing procedures of replacing the unnecessary stroke data input by the user and each constituent element of the closest character by setting restrictions such as not arranging the constituent element of "partial" in the constituent element of "side".

[0062] In step S207, for example, when the constituent element selected by the user's stroke data is [partial] and the constituent element selected for the closest character [qia] is [he] on the [side] side, it is judged that it can be omitted (step S207 "yes"), and the control unit 21 regards the process for the constituent element selected at this time as omitted, returns to the process from step S20, and selects the constituent element of the next closest character shape. ​In addition, in step S207, when it is determined that the selected component cannot be omitted (step S207 "No"), the control unit 21 replaces the component on the [partial] side selected with the stroke data input by the user, for example, with the same component [three dots of water] that is closest to the character [洽] (step S208). Then, the Chinese character candidate based on the replaced component is saved as corrected stroke data (step S209).

[0064] The control unit 21 confirms by determining whether the selection of all the closest components has ended: there are unselected components of the closest characters, and it is not after all the closest components have been selected (step S210).

[0065] In step S210, when it is determined that the selection of all the closest components has not ended (step S210 "No"), the process from step S206 is returned, and the process corresponding to other components of the closest characters is executed.

[0066] In this way, for the component on the [side] of the stroke data input by the user, the process of replacing it with the component [合] on the [side] of the closest character [洽] and setting it as corrected stroke data is executed.

[0067] In step S210, when it is determined that the selection of all the components of the closest characters has ended and there are no unselected components of the closest characters (step S210 "Yes"), the control unit 21 confirms by determining whether the selection of all the components of the stroke data input by the user has ended: there are unselected components, and it is not after all the components have been selected (step S211).

[0068] In step S211, when it is determined that the selection of all the components of the stroke data input by the user has not ended and there are unselected components (step S211 "No"), the control unit 21 returns the process from step S205 and performs the same process on other components of the stroke data input by the user.

[0069] In addition, in step S211, when it is determined that the selection of all the components of the stroke data input by the user has ended and there are no unselected components (step S211 "Yes"), the control unit 21 confirms by determining whether the selection of all the radical-type candidates has ended: there are unselected radical-type candidates, and it is not after all the candidates have been selected (step S212).

[0070] In step S212, when it is determined that the selection of all the radical-type candidates has not ended (step S212 "No"), the control unit 21 returns the process from step S203 and performs the same process based on the unselected radical-type candidates.

[0071] In addition, in step S212, when it is determined that the selection of all radical type candidates has ended (step S212 "Yes"), the control unit 21 ends the subroutine related to the correction process of the stroke data of Figure 4 and returns to Figure 3 the main routine of

[0072] As described above, when the radical type determined from the character image of the handwritten input and the radical type of the character selected as being closer to the character shape are the same, the radical type candidate is limited to one.

[0073] On the other hand, when the radical type determined from the character image of the handwritten input and the radical type of the character selected as being closer to the character shape are different, since two radical type candidates are derived, the processes after step S203 are repeatedly executed.

[0074] Figure 11 This shows an example of the corrected stroke data that replaces the stroke data input by the user with the constituent elements of the character that is the closest to the radical type corresponding to the character [Qia] long-pressed by the user in the candidate character display section 44 of the touch panel display section 17. For the stroke data input by the user, Figure 11 (A) of Figure 11 is the Chinese character candidate [Qia] of the selected closest character shape. The part corresponding to its radical Rix1, i.e., [three dots of water], becomes Figure 11 the corrected stroke data 2 shown in (C) of Figure 11 Similarly, the part corresponding to its side Rix2, i.e., [he], becomes

[0075] the corrected stroke data 1 shown in (B) of Figure 3 In this

[0076] After the correction process of the stroke data is executed in step S104, the handwriting recognition process for multiple correction data is sequentially executed, and corrected Chinese character candidates as the recognition results for the multiple correction data are derived (step S1 (05).

[0077] Figure 7 (B) of Figure 7 is a diagram illustrating the state in which the Chinese character candidates are displayed by referring to the stroke data of the handwritten input through the above series of processes from the state where the Chinese character [Qia] is selected by long-pressing with the stylus pen P in (A) of Figure 7In (B) thereof, the center of the correction Chinese character candidate display section 45 is set to the position of the selected Chinese character "[洽]". To the left of it, two Chinese character candidates "[哈]" and "[恰]" corresponding to the side "[合]" are displayed in a list at 45A. Similarly, two Chinese character candidates "[浴]" and "[沿]" are displayed in a list at 45B on the right with the Chinese character "[洽]" in between. The two Chinese character candidates "[哈]" and "[恰]" at 45A on the left are the candidates derived by performing the handwriting recognition process in step S105 on the correction stroke data 1 shown in (C) below. On the other hand, the two Chinese character candidates "[浴]" and "[沿]" at 45B on the right are the candidates derived by performing the handwriting recognition process in step S105 on the correction stroke data 2 shown in (D) below. Figure 6 In this case, it is also possible to visually represent that they are of the same radical type by surrounding Chinese characters of the same radical type, such as "[哈]" and "[恰]", "[浴]" and "[沿]" with frames of the same color for display. Figure 6 Furthermore, not limited to color, by performing distinguishable display so that the type of frame line, display position, etc. can be visually recognized for each same radical type, the user can easily understand the result of character recognition.

[0078] When the user of the electronic dictionary 10 finds the desired Chinese character among the Chinese character candidates displayed in a list, the user selects the desired Chinese character through a touch operation.

[0079] The control unit 21, as described above, waits for an input of any one of these by repeatedly determining whether the operation of long-pressing the one candidate considered to be the closest (step S103), whether a candidate is selected (step S107), and whether an operation to cancel the list display is performed (step S109).

[0080] Therefore, when a user selects a desired character through a touch operation, the control unit 21 determines in step S107 that a candidate is selected (step S107 "Yes"), determines the selected Chinese character candidate, and inputs it following the already input string (step S108), and the above processing ends.

[0081] As described above Figure 3 The (A) of Figure 7 represents the state where the Chinese character candidate "[浴]" at 45B on the right of the correction Chinese character candidate display section 45 has been touched starting from the display state in (A) of Figure 7 . Corresponding to this touch operation, the control unit 21 determines the Chinese character candidate "[浴]" as the character following the string "[日光]" already input in the retrieval string input section 41.

[0082] Therefore, as described above [[ID=…]] Figure 3 [[ID=…]] [[ID=…]]

[0083] [[ID=…]] Figure 8 The (A) of Figure 7 represents the state starting from the display state in (A) of Figure 7 and touching the Chinese character candidate "[浴]" at 45B on the right of the correction Chinese character candidate display section 45. Corresponding to this touch operation, the control unit 21 determines the Chinese character candidate "[浴]" as the character following the string "[日光]" already input in the retrieval string input section 41. Figure 7 The (A) of Figure 7 represents the state starting from the display state in (A) of Figure 7 and touching the Chinese character candidate "[浴]" at 45B on the right of the correction Chinese character candidate display section 45. Corresponding to this touch operation, the control unit 21 determines the Chinese character candidate "[浴]" as the character following the string "[日光]" already input in the retrieval string input section 41.

[0084] Therefore, as described above​ Figure 8 As shown in (B), in the search string input unit 41, the string "nikko shin" including the misrecognized character "shin" in the temporary input state is replaced with the result "nikko yu" determined through correction processing. At the same time, in the handwritten character input unit 43, the display content is cleared to prepare for the next handwritten input after Chinese character determination.

[0085] In this way, when the desired Chinese character is not recognized through handwritten input, for example, by performing an operation such as a long press on other characters considered to be the closest to the recognition result, as an operation different from the normal touch operation, the stroke data input by the user is decomposed by each radical of the Chinese character to obtain corrected stroke data, and character parsing is performed again. For this reason, even when a part of the handwritten Chinese character is input in a somewhat incorrect form, it can be corrected to increase the possibility of obtaining the desired Chinese character, which is beneficial to the efficiency of handwritten input.

[0086] In addition, for example, when selecting a dictionary function such as a Japanese dictionary, for the characters handwritten in the search string input by the user as part of an entry, based on considering the strings already input before and after it, for the characters considered not to be candidates, by excluding them from the candidates, it is possible to avoid displaying a large number of candidates in an excessive list.

[0087] In addition, in Figure 7 (B) and Figure 8 In the display screen of the touch panel type display unit 17 shown in (A), when the desired Chinese character is still not displayed in either of the Chinese character candidates 45A and 45B in the corrected Chinese character candidate display unit 45, select again the character considered to have the closest shape from them and perform a long press operation, thereby repeatedly executing Figure 4 The correction processing of the stroke data shown.

[0088] Furthermore, when the desired Chinese character is not displayed as a candidate even after repeatedly selecting the Chinese characters considered to have a close shape through long press operations several times, consider performing handwritten input itself again in the handwritten character input unit 43. During such repeated handwritten input, by performing a selection process that reduces the appearance rate of the characters that were previously displayed as candidates but not selected as the desired characters, the recognition accuracy can be improved.

[0089] [Effects of the Embodiment]

[0090] According to the present embodiment described in detail above, the content of handwritten input can be effectively utilized, and the selection of the desired character can be accepted from a wide range of candidates.

[0091] Furthermore, particularly in this embodiment, the stroke data input by the user's handwriting is exported as a radical-type candidate, thereby obtaining corrected stroke data. This radical-type candidate is based on the result of analyzing how to segment the constituent elements of each radical by referring to the radical-type table stored in various table storage areas 22a1. By employing this processing method, although other weighted image processing based on learning, machine learning data, etc., in character recognition processing can be performed simultaneously, the burden of circuit processing in the control unit 21 is reduced in any case, thereby obtaining recognition results more quickly.

[0092] Furthermore, although not described in this embodiment, when the user finally selects a candidate character, the selected character and the handwritten input character can be segmented according to their radicals, and the correspondence between each segmented part can be updated and stored as learning data in various table storage areas 22a1. In this case, the various table storage areas 22a1 function as storage units, storing information indicating the correlation between the segments of the handwritten input character and the corresponding parts of the determined character, which is used when the control unit 21 re-performs character recognition. If the handwriting is the same in subsequent handwritten input character recognitions, it can be automatically corrected based on the learned content. The more it is used, the better it learns the user's handwriting habits, and the higher the recognition rate can be.

[0093] Furthermore, the present invention is not limited to the embodiments described above, and various modifications can be made during the implementation phase without departing from its essence. Moreover, the embodiments can be implemented in suitable combinations to achieve combined effects. Furthermore, the invention, which includes various stages in the above embodiments, can yield various inventions through appropriate combinations of the disclosed constituent elements. For example, if the problem described in the section on the problem to be solved by the invention can be solved even if several constituent elements are deleted from all the constituent elements shown in the embodiments, and the effects described in the section on the effects of the invention can also be obtained, then the structure with the deleted constituent elements can also be extracted as an invention.

[0094] Furthermore, the aforementioned programs can be stored as executable programs on non-temporary computer-readable recording media such as memory cards (ROM cards, RAM cards, etc.), disks (floppy disks, hard disks, etc.), optical discs (CD-ROMs, DVDs, etc.), and semiconductor memories, and distributed accordingly. Moreover, the control unit (CPU) of the electronic device reads the programs recorded on the recording media into the storage device, and by controlling the actions using the read programs, the various functions described in the above embodiments can be realized.

Claims

1. A character recognition device, characterized in that, have: The input unit accepts a character image of the first character input by handwriting; The character derivation unit performs recognition processing on the character image received in the input unit and derives candidate characters in descending order of their similarity to the character image. and The display control unit displays the candidate characters exported from the character export unit as candidate characters for the user's desired character. The input unit can accept a specification operation that derives the character desired by the user from the candidate characters displayed by the display control unit. The character exporting unit uses the correlation with the character image received in the input unit to export the second character. When the character derivation unit receives the specified operation from the input unit, it acquires corrected stroke data of the stroke data of the character image of the first character based on the candidate character shape selected with the highest similarity to the stroke data of the character image received by the input unit, performs recognition processing on the character image based on the corrected stroke data, and derives the second character according to each part segmented according to the density of the character image and the given correlation with the part corresponding to the first character. The display control unit causes the second character exported from the character export unit to be displayed as an additional candidate character expected by the user.

2. The character recognition device according to claim 1, characterized in that, The character derivation unit, corresponding to the regions segmented based on the density of the character image, derives a plurality of second characters having a given association with the first character. The display control unit distinguishes and displays the exported plurality of second characters in relation to the parts segmented according to the density of the character image.

3. The character recognition device according to claim 1 or 2, characterized in that, When the character image received by the input unit is input following a completed string, the character exporting unit exports the second character based on its association with the completed string.

4. The character recognition device according to claim 1 or 2, characterized in that, The input unit can accept a confirmation operation for the second character, which is to be displayed as an additional candidate character. The character recognition device also includes: The storage unit stores information indicating the correlation between the parts segmented according to the density of the character image and the parts corresponding to the second character that received the determination operation, which is used by the character exporting unit to export the second character.

5. A character recognition method, characterized in that, A character recognition method using a device that possesses the following elements: The input unit accepts a character image of the first character input by handwriting; The character derivation unit performs recognition processing on the character image received in the input unit and derives candidate characters in descending order of their similarity to the character image. and The display control unit displays the candidate characters exported from the character export unit as candidate characters for the user's desired character. In the character recognition method, the input unit can accept a specification operation derived from the candidate characters displayed by the display control unit, specifying the character desired by the user. The character exporting unit uses the correlation with the character image received in the input unit to export the second character. When the character derivation unit receives the specified operation from the input unit, it acquires corrected stroke data of the stroke data of the character image of the first character based on the candidate character shape selected with the highest similarity to the stroke data of the character image received by the input unit, performs recognition processing on the character image based on the corrected stroke data, and derives the second character according to each part segmented according to the density of the character image and the given correlation with the part corresponding to the first character. The display control unit causes the second character exported from the character export unit to be displayed as an additional candidate character expected by the user.

6. A recording medium, characterized in that, It is a non-transitory computer-readable recording medium containing a program, and the computer built into the device that has the following elements executes the program: The input unit accepts a character image of the first character input by handwriting; The character derivation unit performs recognition processing on the character image received in the input unit and derives candidate characters in descending order of their similarity to the character image. and The display control unit displays the candidate characters exported from the character export unit as candidate characters for the user's desired character. The program enables the computer to function as follows: The input unit can accept a specification operation that derives the character desired by the user from the candidate characters displayed by the display control unit. The character exporting unit uses the correlation with the character image received in the input unit to export the second character. When the character derivation unit receives the specified operation from the input unit, it acquires corrected stroke data of the stroke data of the character image of the first character based on the candidate character shape selected with the highest similarity to the stroke data of the character image received by the input unit, performs recognition processing on the character image based on the corrected stroke data, and derives the second character according to each part segmented according to the density of the character image and the given correlation with the part corresponding to the first character. The display control unit causes the second character exported from the character export unit to be displayed as an additional candidate character expected by the user.

Images