JP2005078134A - Character recognition device, character recognition method, program and storage medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a character recognition device and a character recognition method, capable of smoothly deciding whether a registered character pattern is proper or not, and efficiently enriching contents of a user dictionary. <P>SOLUTION: A marker (a figure of a black triangle) is additionally inscribed beside a recognition candidate by a character pattern inside the user dictionary among recognition candidates. When a user touches the marker by a pen, a user dictionary management menu is displayed. When touching 'effectiveness down (a downward arrow symbol)' or 'effectiveness up (an upward arrow symbol)' inside the menu by the pen, effectiveness of the character pattern that is the cause of the recognition candidate is increased or reduced. When touching 'delete' by the pen, the character pattern is deleted from the user dictionary. When touching 'image display' by the pen, handwriting of the character pattern is displayed. In reference to the handwriting, the user can easily decide the effectiveness of the character pattern. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

The present invention relates to a character recognition apparatus and a character recognition method for recognizing handwritten character patterns, and is particularly suitable for use in creating a recognition dictionary from user input patterns.

  With the spread of computer systems in recent years, computer systems that emphasize user interface functions have been developed and commercialized. For example, in order to simplify character input, a computer system capable of inputting character information with handwritten characters instead of a conventional keyboard has been commercialized. Such a computer system includes a so-called character recognition device that recognizes predetermined character information from handwriting information handwritten by the user.

  In such a character recognition device, the feature amount of the handwritten character and the feature amount of the recognition dictionary are collated, a character having a high similarity between the two is searched from registered characters in the recognition dictionary, and the search result is presented to the user as a character candidate. The process of doing is performed. Here, the recognition dictionary is usually composed of a standard dictionary and a user dictionary, and among these, the standard dictionary collects, as a sample, feature quantities when an unspecified number of people have written the same character in advance. Created by statistically processing samples. On the other hand, the user dictionary extracts feature amounts from input patterns actually written by the user using the character recognition device, and sequentially stores the feature amounts in association with registered characters designated by the user. . As such user dictionaries are enriched, the probability that correct characters will be presented as character candidates also increases depending on the user's writing, so that it is possible to shorten the time required to properly input the correct character desired by the user.

Here, a function for optimizing the user dictionary is added to the character recognition device. For example, in Patent Document 1 below, when a character pattern is automatically registered in the user dictionary, a handwriting of the character pattern to be registered is displayed in advance, and the user is confirmed as to whether or not registration is appropriate.
JP-A-6-162267

  According to the method of the above-mentioned patent document 1, a character pattern can be registered on the user's initiative, and thus a user who is more adapted to an individual user than a case in which a character pattern is randomly registered at the judgment of the device. You can build a dictionary.

  However, since the method of Patent Document 1 relates to processing at the time of registration, once it has been registered, its suitability cannot be determined any longer. However, in actual use, it is often the case that character patterns judged to be valid at the time of registration are an obstacle to the selection of correct characters, and it is important to organize character patterns after registration rather than at the time of registration. Often.

Therefore, an object of the present invention is to provide a character recognition device and a character recognition method that can smoothly determine the suitability of a registered character pattern and can efficiently enhance the contents of a user dictionary.

  In view of the above problems, the present invention has the following features.

  The invention of claim 1 is a handwritten character input means for inputting a handwritten character pattern, a standard dictionary for storing character pattern feature values in association with registered characters, character pattern feature values, and effectiveness of the character patterns. A user dictionary stored in association with registered characters; a recognition candidate generating unit that generates a recognition candidate by comparing the feature amount of the character pattern input by handwriting with the feature amount of the standard dictionary and the user dictionary; and the recognition Recognition candidate display means for displaying the recognition candidates generated by the candidate generation means, and information on the character patterns stored in the user dictionary when the recognition candidates are based on the character patterns in the user dictionary A change screen display means for displaying an input screen for changing the character string in association with the candidate character, and the user according to the input to the input screen. A character recognition apparatus characterized by having a user dictionary changing means for changing the information of the character patterns stored in the dictionary.

  According to the present invention, it is possible to adjust the character pattern information that is the basis of the recognition candidate in association with the display of the recognition candidate. Therefore, the character pattern information can be adjusted while determining the influence of the character pattern on the recognition candidate. Therefore, the contents of the user dictionary can be effectively enhanced.

  Here, the display process in the change screen display means displays a marker for identifying a recognition candidate based on the character pattern in the user dictionary in association with the recognition candidate, as described in claim 2. An input screen for changing the information of the character pattern in accordance with an instruction input to the marker can be configured to be displayed. According to the present invention, since the information on the character pattern can be changed by inputting the marker, the information on the character pattern can be appropriately changed as necessary.

  The display processing in the change screen display means is performed when the recognition candidate based on the character pattern in the user dictionary exists in a candidate rank higher than the correct character as described in claim 3. The input screen may be displayed in association with the. According to the present invention, it is possible to present a user with a character pattern that is an obstacle to a correct character, and to prompt the user to change information regarding the character pattern. Accordingly, it is possible to appropriately change the information regarding the character pattern while determining the magnitude of the obstacle of the character pattern with respect to the correct character, and as a result, it is possible to effectively enhance the contents of the user dictionary.

  In addition, the display process in the change screen display means may be configured to associate the recognition candidate with the recognition candidate when the recognition candidate for the correct character is a recognition candidate based on the character pattern in the user dictionary. It can also be configured to display an input screen. According to this invention, it is possible to present to the user a character pattern that has contributed to the correct character, and to prompt the user to change information related to the character pattern. Therefore, it is possible to appropriately change the information regarding the character pattern while determining the degree of contribution to the correct character, and as a result, it is possible to effectively enrich the contents of the user dictionary.

  Furthermore, as described in claim 5, if the handwriting of the character pattern is displayed when the change screen is displayed, the user can actually write the handwriting of the character pattern that is the basis of the recognition candidate. It is possible to confirm, and according to the confirmation result, it is possible to determine whether the information regarding the character pattern should be changed.

  In the above, as described in claim 6, the input screen includes at least an input item for changing the effectiveness of the character pattern and / or an input item for deleting the character pattern from the user dictionary. Should be included.

  The inventions according to claims 7 to 12 are obtained by extracting the above invention as a method invention. The invention of claim 13 is a program for giving a function of each processing in the character recognition method according to any one of claims 7 to 12 to the computer, and the invention of claim 14 is the claim. A storage medium storing a program for giving a function of each process in the character recognition method according to any one of 7 to 12 to a computer.

  The features of the present invention and the technical significance and effects thereof will become more apparent from the following description of embodiments.

However, the following embodiment is only one embodiment of the present invention, and the meaning of the present invention or the terms of each constituent element is limited to those described in the following embodiment. is not.

  FIG. 1 shows a functional block diagram of the character recognition apparatus according to the embodiment.

  As illustrated, the character recognition device includes a handwriting input unit 1, an input processing unit 2, a control unit 3, a recognition candidate memory 4, a character candidate selection unit 5, a display control unit 6, a display unit 7, a feature amount extraction unit 8, It comprises a character recognition unit 9, a user dictionary management control unit 10, a recognition result memory 11, and a data bus 12.

  The handwriting input unit 1 is configured by a display input unit such as a tablet, and outputs coordinate information of a position touched with a pen to the input processing unit 2. The input processing unit 2 determines a pen-touched area based on the coordinate information input from the handwriting input unit 1 and outputs information corresponding to each area to a corresponding function unit. For example, if the pen touch area is a character entry frame, the coordinate information input from the handwriting input unit 1 is output to the feature amount extraction unit 8. If the pen touch area is a command input area for executing predetermined processing (starting character recognition, selecting a recognition candidate, displaying the next recognition candidate, displaying the previous recognition candidate, etc.), the command input The input command is output to a functional unit that needs to receive

  The control unit 3 controls each unit when executing the function of the character recognition device. The recognition candidate memory 4 stores the recognition candidates generated by the character recognition unit 9. The recognition candidate generation process and the content of information stored in the recognition candidate memory 4 will be described in detail later.

  The character candidate selection unit 5 displays the highest recognition candidate among the recognition candidates stored in the recognition candidate memory 4 on the display unit 7 as a recognition result, and when the user selects a desired character from the recognition candidates. The selected recognition candidate is displayed on the display unit 7 as a recognition result. A specific display form will be described in detail later.

  The display control unit 6 displays a predetermined display screen on the display unit 7 in response to a command from the control unit 3. The display unit 7 is configured by a display input unit such as a tablet, and displays a character entry frame, various command input buttons, and a handwriting of characters input by the user according to control by the display control unit 6. The display screen displayed at the time of character recognition will be described in detail later.

  The feature amount extraction unit 8 processes the handwriting coordinate information received from the input processing unit 2 to extract the feature amount of the input character, and outputs this to the character recognition unit 9. The character recognition unit 9 compares the feature amount received from the feature amount extraction unit 8 with the feature amount in the standard dictionary 9a and the user dictionary 9b to calculate the similarity, and the characters from the higher similarity to the predetermined rank The group is output as a recognition candidate to the recognition candidate memory 4.

  The user dictionary management control unit 10 updates the user dictionary 9b in response to an input command to a user dictionary management menu (described later) displayed during the character recognition process. Details of the update process will be described later. The recognition result memory 11 sequentially stores the correct characters determined by the character recognition process.

  FIG. 2 shows the configuration of information stored in the standard dictionary 9a and the user dictionary 9b. As illustrated, the standard dictionary stores character pattern feature values in association with recognition results. Here, each character pattern has a pattern No. Specified by. The user dictionary stores character pattern feature amounts, pattern information, and validity levels in association with recognition results. Here, the pattern information is handwriting information (coordinate information) of the character pattern. The validity is set to 1.00 at the time of character pattern registration, and is changed by subsequent adjustment processing (described later). Each character pattern has a pattern no. Specified by.

  Next, a recognition candidate generation process and information storage process for the recognition candidate memory will be described with reference to FIG.

  For example, when an input character as shown in the figure is input by writing, the character recognition unit 9 performs the feature amount of the input character and the feature amount of the character pattern registered in the standard dictionary 9a and the user dictionary 9b. Are compared in order, and the similarity S between them is calculated. This figure shows the calculation result of the similarity S between the character pattern registered in association with the recognition result “river”. Here, for the character pattern in the user dictionary 9b, the similarity S 'is calculated by multiplying the calculated similarity S by the effectiveness of each character pattern. Then, the similarity S of each character pattern in the standard dictionary 9a and the similarity S 'of each character pattern in the user dictionary 9b are compared for each recognition result, and the highest similarity is used as the similarity of the recognition result. In the case of the figure, the pattern No. in the user dictionary 9b. Since the similarity S ′ = 0.92 of 002 is the highest, this similarity is set as the similarity of the recognition result.

  The character recognition unit 9 temporarily stores information about each recognition result in the built-in memory. FIG. 4A shows the contents of temporarily stored information. As shown in the figure, the built-in memory has a recognition result, a dictionary flag, and a pattern number. Are stored in association with each other. Here, the dictionary flag indicates whether the similarity of each recognition result is based on the character pattern of the standard dictionary 9a (flag = 0) or the character pattern of the user dictionary 9b (flag = 1). is there.

  In the built-in memory, for example, information on 10 sets of recognition results with higher similarity is stored as needed. That is, when the similarity is determined for a new recognition result by the above processing, the similarity is larger than the lowest similarity among the similarities of the 10 recognition results already stored in the built-in memory. If it is large, the information on the new recognition result is replaced with the recognition result having the lowest similarity and stored in the built-in memory. If it is small, the information on the new recognition result is stored in the built-in memory. Is processed so that it is not stored.

  When the processing is completed for all the dictionary information in this way, the recognition candidates in the built-in memory are rearranged according to the order of similarity to generate recognition candidate information as shown in FIG. This is transferred to and stored in the recognition candidate memory 4.

  Next, a display screen displayed on the display unit 7 will be described with reference to FIG.

  After writing a character in the character entry box and touching the recognition button with a pen, the character recognition process is executed as described above to generate a recognition candidate. Such recognition candidates are displayed in the character candidate display frame in descending order. Further, among the recognition candidates displayed in the character candidate display frame, markers (black triangles in the figure) are superimposed and displayed on the recognition candidates based on the character pattern in the user dictionary 9b. That is, the display control unit 6 refers to the recognition candidate memory 4 (see FIG. 3B), identifies a recognition candidate whose dictionary flag is “1”, and adds a marker to the recognition candidate.

  In addition, among the recognition candidates, the candidate character with the highest candidate rank is displayed in the recognition result display frame. Further, the handwriting of the input character is displayed as it is in the character entry frame. FIG. 4A shows the display state at this time. In the display example of the figure, the top five recognition candidates are displayed in the character candidate display frame.

  In this display state, when the OK button is touched with a pen, the recognition result is fixed to “river”, and the handwriting in the character entry frame is deleted. The confirmed recognition result is stored in the recognition result memory 11. When the feed button (>>) is touched with a pen, the following five recognition candidates are displayed in the character candidate display frame. Thereafter, when the return button (<<) is touched with a pen, the previously displayed recognition candidates are displayed again in the character candidate display frame.

  When “m” displayed in the character candidate display frame is pen-touched in the state of FIG. 10A, the display in the recognition result display frame is changed from “river” to “m” as shown in FIG. Be changed. Thereafter, when the OK button is touched with a pen, the recognition result is fixed to “m”, and the handwriting in the character entry frame is deleted. The confirmed recognition result is stored in the recognition result memory 11.

  In the state shown in FIG. 5A, when the marker is touched with a pen, a user dictionary management menu shown in FIG. 5A is displayed. That is, the display control unit 6 receives the information of pen touching the marker from the input processing unit 2 and superimposes the user dictionary management menu on the display screen. The user can adjust the effectiveness of the character pattern by selecting a predetermined item in the displayed menu.

  For example, when the “validity ↑” in the menu displayed in FIG. 5A is pen-touched, the effectiveness of the character pattern that is the basis of the recognition candidate “river” is increased by a predetermined value ΔVa. That is, when “validity ↑” is touched with a pen, the user dictionary management control unit 10 refers to the recognition candidate memory 4 (see FIG. 3B), and the character pattern that is the basis of the recognition candidate “river”. Pattern No. Identifies based on In the case of FIG. 3, the character pattern that is the basis of the recognition candidate “river” is the pattern number. 002 character pattern. Thereafter, the user dictionary management control unit 10 specifies the pattern No. specified in the user dictionary 9b. Is increased by the value ΔVa.

  Similarly, when “effectiveness ↓” (effectiveness down) is touched with a pen in the user dictionary management menu of FIG. 5, the effectiveness of the character pattern (pattern No. 002) that is the basis of the recognition candidate “river”. Is reduced by a predetermined value ΔVb. In addition, when “delete” is touched with a pen in the menu, the character pattern (pattern No. 002) that is the basis of the recognition candidate “river” is deleted from the user dictionary 9b. Furthermore, when “CLOSE” is touched with a pen in the menu, the user dictionary management menu is closed and the display state of FIG. 4A is restored.

  In the user dictionary management menu of FIG. 5, when “image display” is touched with a pen, the character pattern (pattern No. 002) that is the basis of the recognition candidate “river” is displayed as shown in FIG. Handwriting is displayed. That is, the display control unit 6 reads the pattern information of the character pattern (pattern No. 002) that is the basis of the recognition candidate “river” from the user dictionary 9b, generates a handwriting image based on the pattern information, This is displayed on the display unit 7.

  As a result, the user can actually confirm the handwriting of the character pattern that is the basis of the recognition candidate, and the effectiveness of the character pattern should be adjusted according to the confirmation result, or the character It can be determined whether the pattern should be registered. If it is necessary to adjust the effectiveness or delete the character pattern, touch “Return” to return to the menu display of FIG. 5A, and input the adjustment command for the effectiveness or delete the character pattern as appropriate. I do.

  FIG. 6 shows a flowchart of processing when the user dictionary 9b is adjusted.

  When the handwritten character is input and the character recognition process is executed by the user, the recognition candidates are displayed in the character candidate display frame as described above (S101). In addition, among the displayed recognition candidates, the above marker is superimposed on the recognition candidates based on the character pattern in the user dictionary 9b (S102). Thereafter, when the user confirms a predetermined recognition candidate as a correct character (S103), the adjustment process of the user dictionary 9b is terminated. On the other hand, when the user pen-touches the marker displayed in the character candidate display frame (S104), the user dictionary management menu is displayed in an overlapping manner (S105).

  When a predetermined item in the displayed menu is touched with a pen (S106), processing according to the item touched with the pen is executed as described above (S107). When the processing is completed, the user dictionary management menu is closed and the recognition candidates and markers are displayed again (S108). At this time, if a character pattern is deleted from the user dictionary 9b in S107, it is preferable not to add a marker to a recognition candidate based on the character pattern.

  Thereafter, the process returns to S103, and the same process as described above is executed again. This process is repeated until the correct character confirmation process is completed (S103). When the correct character confirmation process is completed, the handwriting in the character entry frame and the recognition candidate in the character candidate display frame are deleted, and at the same time, the adjustment process of the user dictionary 9b is also terminated.

  According to the present embodiment, when the recognition candidate is displayed, the character pattern that is the basis of the recognition candidate can be adjusted at the same time. Therefore, the character pattern is appropriately adjusted while determining the influence of the character pattern on the recognition candidate. Therefore, the contents of the user dictionary can be effectively enhanced.

  Next, another embodiment according to the present invention will be described.

  In the above embodiment, the marker is added to prompt the user to adjust the character pattern. However, in the present embodiment, the character from the user dictionary 9b is placed higher than the recognition candidate confirmed as the correct character. When there is a recognition candidate by pattern, the user dictionary management menu is automatically displayed to prompt the user to adjust the character pattern. In other words, when there are recognition candidates based on the character pattern from the user dictionary 9b higher than the recognition candidates determined as correct characters, there is a possibility that such character patterns have hindered the selection of correct characters. Therefore, the user is prompted to adjust the effectiveness of the character pattern.

  In the present embodiment, as shown in FIG. 7A, no marker is added in the character candidate display frame, and only the recognition candidates are displayed. Therefore, from such a display, the user cannot know which recognition candidate is based on the character pattern from the user dictionary 9b. Here, assuming that “river” is a character pattern from the user dictionary among the recognition candidates, the user sets the recognition candidate “m” ranked second in the candidate ranking as the correct character in the display of FIG. When it is determined, a user dictionary management menu as shown in FIG. 5B is displayed for the higher recognition candidate “river”.

  Specifically, the display control unit 6 refers to the recognition candidate memory 4 when correct character determination processing is performed, and determines whether or not a recognition candidate having a dictionary flag “1” exists above the correct character. judge. If there is a recognition candidate whose dictionary flag is “1”, the recognition candidate is selected as a display target of the user dictionary management menu, and a user dictionary management menu display screen corresponding to the recognition candidate is generated. As shown in FIG. 7B, the images are displayed in an overlapping manner on the display screen.

  For example, when the contents of the recognition candidate memory 4 are as shown in FIGS. 8A and 8B, it is assumed that the recognition candidates indicated by the arrows in FIG. 8 are confirmed as correct characters. In this case, in FIG. 9A, since the recognition candidate of the dictionary flag “1” does not exist above the correct character, the display control unit 6 does not display the user dictionary management menu. On the other hand, in the case of FIG. 5B, since the recognition candidate (first candidate ranking) of the dictionary flag “1” exists above the correct character, the display control unit 6 relates to this recognition candidate. Displays the user dictionary management menu. That is, the display control unit 6 displays the pattern No. The pattern information of the character pattern 002 is extracted from the user dictionary 9b to generate a handwriting image. Then, the user dictionary management menu shown in FIG. 7B is displayed on the screen so that the generated handwriting image is displayed at a predetermined position.

  The user dictionary management menu shown in FIG. 7 displays handwritten images at the same time. However, as in FIG. 5 shown in the first embodiment, the item “image display” is added to the user dictionary management menu. A handwriting image may be displayed in response to selection of this item.

  The user can adjust the effectiveness of the character pattern by selecting a predetermined item in the displayed menu.

  For example, when the “validity ↓” in the menu displayed in FIG. 7B is pen-touched, the effectiveness of the character pattern that is the basis of the recognition candidate “river” is reduced by a predetermined value ΔVb. That is, when “effectiveness ↓” is touched with a pen, the user dictionary management control unit 10 refers to the recognition candidate memory 4 (see FIG. 8), and determines the character pattern that is the basis of the recognition candidate “river” as a pattern number. . Identifies based on In the case of FIG. 8, the character pattern that is the basis of the recognition candidate “river” is the pattern number. 002 character pattern. Thereafter, the user dictionary management control unit 10 specifies the pattern No. specified in the user dictionary 9b. Is reduced by the value ΔVb.

  In addition, when “delete” is touched with a pen in the menu, the character pattern (pattern No. 002) that is the basis of the recognition candidate “river” is deleted from the user dictionary 9b. Furthermore, when “CLOSE” is touched with a pen in the menu, the user dictionary management menu is closed and the display state of FIG. 7A is restored.

  FIG. 9 shows a flowchart of processing when the user dictionary 9b is adjusted.

  When handwritten characters are input by the user and character recognition processing is executed, recognition candidates are displayed in the character candidate display frame as described above (S201). Next, when the correct character determination process is performed by the user (S202), it is determined whether there is a recognition candidate (M) based on the character pattern from the user dictionary above the correct character (S203), and the recognition candidate (M ) Exists, all recognition candidates (M) are selected as display targets of the user dictionary management menu.

  Thus, when a display target is selected, first, a display image of the user dictionary management menu is generated for the recognition candidate (M) with the highest candidate rank, and is superimposed on the display screen (S204). When a predetermined item is touched with a pen in the displayed menu (S205), processing according to the pen-touched item is executed as described above (S206). Then, when this process is completed, the user dictionary management menu for the recognition candidate (M) is closed. Thereafter, it is determined whether the user dictionary management menu is displayed for all recognition candidates (M) selected in S203 (S207). If there is a recognition candidate (M) for which no menu has been displayed, the next candidate is displayed. The processing from S204 onward is executed for the rank recognition candidate (M).

  This process is repeated until the user dictionary management menu is displayed for all recognition candidates (M) (S207). When the processing for all the recognition candidates (M) is completed, the handwriting in the character entry frame and the recognition candidates in the character candidate display frame are deleted, and at the same time, the adjustment process of the user dictionary 9b is also ended.

  According to the present embodiment, the character pattern that is an obstacle to the correct character is presented to the user, and the adjustment process of the character pattern is urged to the user. While making the determination, the character pattern can be adjusted appropriately, and the contents of the user dictionary can be effectively enhanced.

  By the way, in the said embodiment, the character pattern which became the obstacle of the correct character was shown to the user, and the adjustment process of the said character pattern was prompted, but the correct character is based on the character pattern from the user dictionary 9b. In such a case, the user may be prompted to perform adjustment processing of the character pattern (for example, processing for further increasing the effectiveness). For example, in the display state of FIG. 10A, when the recognition candidate “river” is a character pattern from the user dictionary 9b, the user confirms the recognition candidate “river” as a correct character, The user dictionary management menu shown in FIG.

  Specifically, the display control unit 6 refers to the recognition candidate memory 4 and determines whether the dictionary flag of the correct character is “1” when the correct character is determined. If the dictionary flag is “1”, the recognition candidate is selected as a display target of the user dictionary management menu, and a user dictionary management menu display screen corresponding to the recognition candidate is generated. As shown in b), it is displayed in an overlapping manner on the display screen.

  For example, when the contents of the recognition candidate memory 4 are as shown in FIGS. 11A and 11B, it is assumed that the recognition candidates indicated by the arrows in FIG. 11 are confirmed as correct characters. In this case, since the dictionary flag of the correct character is not “1” in FIG. 5B, the display control unit 6 does not display the user dictionary management menu. On the other hand, in the case of FIG. 5A, since the correct character dictionary flag is “1”, the display control unit 6 displays the user dictionary management menu for this recognition candidate. That is, the display control unit 6 displays the pattern No. The pattern information of the character pattern 002 is extracted from the user dictionary 9b to generate a handwriting image. Then, the user dictionary management menu shown in FIG. 10B is displayed on the screen so that the generated handwriting image is displayed at a predetermined position.

  FIG. 12 shows a flowchart of processing when the user dictionary 9b is adjusted.

  When handwritten characters are input by the user and character recognition processing is executed, recognition candidates are displayed in the character candidate display frame as described above (S301). Next, when the correct character determination process is performed by the user (S302), it is determined whether the correct character is a character pattern from the user dictionary (S303). If YES, the correct character is the user dictionary. Selected as a display target of the management menu.

  Then, a display image of the user dictionary management menu is generated for the correct answer character, and is superimposed on the display screen (S304). When a predetermined item in the displayed menu is touched with a pen (S305), processing according to the item touched with the pen is executed as described above (S306). Then, when this process is completed, the user dictionary management menu for the recognition candidate (M) is closed. At the same time, the handwriting in the character entry frame and the recognition candidates in the character candidate display frame are erased, and the adjustment process of the user dictionary 9b is also terminated.

  According to the present embodiment, the character pattern that contributed to the correct character is presented to the user, and the user is prompted to adjust the character pattern. Therefore, while determining the degree of contribution to the correct character, The character pattern can be adjusted appropriately, and the contents of the user dictionary can be effectively enhanced.

  In each of the above embodiments, the new character pattern is registered in the user dictionary 9b. For example, when the recognition candidate confirmed as the correct character is a character pattern from the standard dictionary 9a, (Amount, pattern information) may be automatically registered in the user dictionary 9b in association with the recognition result of the correct character. This increases the possibility that the recognition result based on the new character pattern will be set as a recognition candidate when the same written input is made next time, and thus the recognition reflecting the more recent character input tendency. Candidates can be provided to users.

  The embodiment according to the present invention has been described above. However, the present invention is not limited to the embodiment, and it is needless to say that other various modifications are possible. For example, in the above embodiment, the user dictionary management menu is configured as shown in FIGS. 5, 7, and 10, but the user dictionary management menu may be configured in other display formats.

  Needless to say, the character recognition apparatus described in the above embodiment can be realized by installing all or a part of a program for executing the function and various databases in a general-purpose computer apparatus or the like. The installation of the program or the like can be realized by downloading via a network such as the Internet in addition to a configuration in which a storage medium such as an optical disk storing the program or the like is mounted on the computer device.

The embodiments of the present invention can be appropriately modified in various ways within the scope of the technical idea shown in the claims.

Functional block diagram of a character recognition device according to an embodiment The figure which shows the content of the standard dictionary and user dictionary which concern on embodiment The figure explaining the recognition candidate production | generation process which concerns on embodiment The figure which shows the display form of the display part which concerns on embodiment The figure which shows the display form of the display part which concerns on embodiment User dictionary adjustment process flow according to the embodiment The figure which shows the display form of the display part which concerns on 2nd Embodiment The figure which shows the content of the recognition candidate memory which concerns on 2nd Embodiment User dictionary adjustment process flow according to the second embodiment The figure which shows the display form of the display part which concerns on 3rd Embodiment. The figure which shows the content of the recognition candidate memory which concerns on 3rd Embodiment User dictionary adjustment process flow according to the third embodiment

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Handwriting input part 2 Input processing part 4 Recognition candidate memory 6 Display control part 7 Display part 8 Feature-value extraction part 9 Character recognition part 9a Standard dictionary 9b User dictionary 10 User dictionary management control part

Claims (14)

A handwritten character input means for inputting a handwritten character pattern;
A standard dictionary that stores character pattern feature values in association with registered characters;
A user dictionary that stores character pattern feature quantities and the effectiveness of the character patterns in association with registered characters;
A recognition candidate generating unit that generates a recognition candidate by comparing the feature amount of the character pattern input by handwriting with the feature amount of the standard dictionary and the user dictionary;
Recognition candidate display means for displaying the recognition candidates generated by the recognition candidate generation means;
When the recognition candidate is based on a character pattern in the user dictionary, an input screen for changing information on the character pattern stored in the user dictionary is displayed in association with the candidate character. Screen display means;
User dictionary changing means for changing information of the character pattern stored in the user dictionary in response to an input to the input screen;
A character recognition device. In claim 1,
The change screen display means displays a marker for identifying a recognition candidate based on a character pattern in the user dictionary among the recognition candidates displayed on the recognition candidate display means in association with the recognition candidate, Display an input screen for changing the information of the character pattern in response to an instruction input to the marker;
A character recognition device. In claim 1,
The change screen display means displays the input screen in association with the recognition candidate when there is a recognition candidate based on the character pattern in the user dictionary in a candidate rank higher than the correct character.
A character recognition device. In claim 1 or 3,
When the correct character recognition candidate is a recognition candidate based on a character pattern in the user dictionary, the change screen display means displays the input screen in association with the recognition candidate.
A character recognition device. In any of claims 1 to 4,
In the user dictionary, in addition to the feature amount of the character pattern and the effectiveness of the character pattern, handwriting information of the character pattern is stored in association with the registered character,
The change screen display means displays the handwriting of the character pattern simultaneously with the input screen or in response to an instruction input to the input screen.
A character recognition device. In any of claims 1 to 5,
The input screen includes an input item for changing the effectiveness of the character pattern and / or an input item for deleting the character pattern from the user dictionary.
A character recognition device. Characters for character recognition using a standard dictionary that stores character pattern feature values in association with registered characters, and a user dictionary that stores character pattern feature values and validity of the character patterns in association with registered characters A recognition method,
A recognition candidate generation process for generating a recognition candidate by comparing the feature amount of the character pattern input by handwriting with the feature amount of the standard dictionary and the user dictionary;
A recognition candidate display process for displaying the recognition candidates generated by the recognition candidate generation means;
When the recognition candidate is based on a character pattern in the user dictionary, an input screen for changing information on the character pattern stored in the user dictionary is displayed in association with the candidate character. Screen display processing,
A user dictionary change process for changing information on the character pattern stored in the user dictionary in response to an input on the input screen.
A character recognition method characterized by the above. In claim 7,
The change screen display process displays a marker for identifying a recognition candidate based on a character pattern in the user dictionary among the recognition candidates displayed in the recognition candidate display process in association with the recognition candidate. Display an input screen for changing the information of the character pattern in response to an instruction input to the marker;
A character recognition method characterized by the above. In claim 7,
The change screen display process displays the input screen in association with the recognition candidate when there is a recognition candidate based on the character pattern in the user dictionary in a candidate rank higher than the correct character.
A character recognition method characterized by the above. In claim 7 or 9,
The change screen display process displays the input screen in association with the recognition candidate when the correct character recognition candidate is a recognition candidate based on a character pattern in the user dictionary.
A character recognition method characterized by the above. In any of claims 7 to 10,
In the user dictionary, in addition to the feature amount of the character pattern and the effectiveness of the character pattern, handwriting information of the character pattern is stored in association with the registered character,
The change screen display process displays the handwriting of the character pattern simultaneously with the input screen or in response to an instruction input to the input screen.
A character recognition method characterized by the above. In any of claims 7 to 11,
The input screen includes an input item for changing the effectiveness of the character pattern and / or an input item for deleting the character pattern from the user dictionary.
A character recognition method characterized by the above.   13. A program for giving a function of each process in the character recognition method according to claim 7 to a computer. A storage medium storing a program for giving a function of each process in the character recognition method according to any one of claims 7 to 12 to a computer.

Images