JP2014056389A - Character recognition device, character recognition method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a character recognition device capable of easily notifying a user that a context correction is made on an input character string to prompt the user to give a correction instruction on the character string after the context correction if necessary.SOLUTION: A recognition character replacement determination section 23 determines whether a character string as a character recognition result is subjected to a context correction (replacement) by a context correction processing section 20. A first candidate periphery display processing section 24 displays a first candidate character string in the periphery (for example, immediately above) of the character string as the character recognition result. When the first candidate is selected, a recognition character correction section 22 corrects the character string after the context correction to the first candidate character string. A prediction conversion candidate extraction section 18 updates at least one prediction conversion candidate including the character string after being subjected to the context correction to at least one prediction conversion candidate including the first candidate character string according to the correction to the first candidate character string.

Description

  The present invention relates to a character recognition device, a character recognition method, and a program for recognizing handwritten characters.

  For example, Patent Documents 1 and 2 are known as prior arts for recognizing a character handwritten on a touch panel by a user via an instruction medium such as a finger or a stylus pen.

  The information processing apparatus shown in Patent Literature 1 detects input handwriting and performs handwriting recognition processing, and stores handwriting information and a plurality of recognition results as handwriting recognition processing results in association with each other. The information processing apparatus concatenates handwriting information corresponding to a plurality of recognition results and a plurality of recognition results corresponding to the preceding and succeeding handwriting information, and re-decides the priority of the recognition candidates based on the context, and performs narrowing down. Thus, one recognition candidate is selected.

  In the data processing device shown in Patent Literature 2, when a handwritten character input frame on which one of a plurality of character candidates is displayed is designated by a pen, a character candidate for a handwritten character in the handwritten character input frame, and deletion of the input frame, Edit commands that can be executed such as rewriting handwritten characters are displayed together. Furthermore, when any of the character candidates is selected, the data processing apparatus changes the character candidate in the handwritten character input frame to the selected character candidate and displays it.

Japanese Patent No. 3351039 JP 2000-099223 A

  However, the information processing apparatus disclosed in Patent Literature 1 narrows down recognition candidates in consideration of the context with the preceding and following handwriting information, even though the recognition result of the handwriting recognition processing for the input handwriting is a correct character string. For this reason, depending on the input handwriting, a recognition candidate not intended by the user may be selected, and a character string handwritten by the user may be erroneously recognized.

  Further, in the data processing device shown in Patent Document 2, when a handwritten input character string is erroneously recognized, the user may not be aware that the user has been erroneously recognized. Even if the user notices that the recognition error has occurred, in order to correct a recognition candidate that has been erroneously recognized, it is necessary to perform a predetermined operation to display a list of recognition candidates and to further select the correct recognition candidate. As a result, the user's operation was complicated.

  In order to solve the above-described conventional problems, the present invention makes it easy to notice that context correction has been performed on an input character string, and if necessary, a correction instruction for the character string after context correction is appropriately issued. An object of the present invention is to provide a character recognition device, a character recognition method, and a program for prompting.

  A character recognition device according to the present invention includes a display unit that displays data on a screen, a touch panel that is provided in correspondence with the screen and receives handwritten input, a character recognition processing unit that recognizes the handwritten handwritten character, Based on the context of the recognized character display processing unit that displays the character recognition result of the character recognition on the screen and the character string before and after the handwriting when the handwritten input handwriting is two or more characters, A context correction processing unit that corrects any character string in the handwriting, a replacement determination unit that determines whether or not any of the character strings has been context corrected, and any of the character strings has been context corrected. A first candidate that, when determined, displays the character recognition result character string before the context correction as a first candidate character string around a display position of the character string of the character correction result subjected to the context correction; table Comprising a processing unit.

  According to the configuration described above, the replacement determination unit determines that context correction has been performed on the input character string, and the character before the context correction is displayed around the display position of the character string of the character recognition result after context correction. Since the character string of the recognition result is displayed as the first candidate character string, it is easy to notice that the context correction for the input character string has been executed, and if necessary, an instruction to correct the character string after the context correction is given. Can prompt you in a timely manner.

  The character recognition method of the present invention is a character recognition method in a character recognition device having a display unit for displaying data on a screen, the character recognition method being provided corresponding to the screen, and receiving the handwriting input. A step of recognizing the handwritten character, a step of displaying the character recognition result of the character recognition on the screen, and a character string before and after the handwriting when the handwritten input handwriting is two or more characters, Based on the context, the step of correcting any character string in the handwriting, the step of determining whether any of the character strings has been context corrected, and the context correction of any of the character strings The character recognition result character string before the context correction is set as a first candidate character string around the display position of the character recognition result character string after the context correction. And a step of displaying Te.

  According to the above-described method, the replacement determination unit determines that context correction has been performed on the input character string, and the character before the context correction is displayed around the display position of the character string as the result of character recognition after context correction. Since the character string of the recognition result is displayed as the first candidate character string, it is easy to notice that the context correction for the input character string has been executed, and if necessary, an instruction to correct the character string after the context correction is given. Can prompt you in a timely manner.

  The program of the present invention is provided in a character recognition device, which is a computer having a display unit for displaying data on a screen, corresponding to the screen, and accepts handwritten input. The step of recognizing, the step of displaying the recognized character recognition result on the screen, and when the handwritten input handwriting is two or more characters, based on the context of the character string before and after the handwriting A step of correcting any character string in the handwriting, a step of determining whether any of the character strings has been context-corrected, and a case of determining that any of the character strings has been context-corrected. In addition, the character recognition result character string before the context correction is displayed as a first candidate character string around the display position of the character recognition result character string after the context correction. A step that causes the execution.

  According to the above-described program, the replacement determination unit determines that context correction has been performed on the input character string, and the character before the context correction is displayed around the display position of the character string of the character recognition result after context correction. Since the character string of the recognition result is displayed as the first candidate character string, it is easy to notice that the context correction for the input character string has been executed, and if necessary, an instruction to correct the character string after the context correction is given. Can prompt you in a timely manner.

  According to the present invention, it is possible to easily recognize that context correction has been performed on an input character string, and to prompt a correction instruction for the character string after context correction as needed.

The block diagram which shows the internal structure of the character input device of this embodiment (A), (B) The figure which shows the display content displayed on the screen of a display apparatus (A)-(C) Screen transition diagram for explaining the operation outline of the character input device of this embodiment in time series (A)-(C) Screen transition diagram for explaining the operation outline of the character input device of this embodiment in time series The flowchart explaining the operation | movement procedure of the character input device of this embodiment. The flowchart explaining the operation | movement procedure of the character input device of this embodiment. The figure which shows the variation of the display position of the character string of a 1st candidate, (A) The example displayed on the upper part of the recognition character string "RO", (B) The example displayed on the lower part of the recognition character string "RO", ( C) Example displayed on the right side of the recognized character string “RO”, (D) Example displayed on the upper right side of the recognized character string “RO” The figure which shows the example in which the character string of the some 1st candidate was displayed on the upper part of the some recognition character string

  Hereinafter, embodiments of a character recognition device among character recognition devices, character recognition methods, and programs according to the present invention will be described with reference to the drawings. A character recognition device according to the present invention is an electronic device including a display unit having a function of displaying data on a screen (display screen), for example, a mobile phone, a smartphone, a tablet terminal, a digital still camera, a PDA (personal digital assistant). Or it is an electronic book terminal. Hereinafter, a character input device (for example, a smartphone) will be described as an example of the character recognition device of each embodiment.

  The present invention can also be expressed as a character input device as a device or a program for operating the character input device as a computer. Furthermore, the present invention can be expressed as a character recognition method including each operation (step) executed by the character input device. That is, the present invention can be expressed in any category of an apparatus, a method, and a program.

  In the following description, when a user's finger or stylus pen handwriting input operation starts or resumes, the user's operation in which the finger or stylus pen contacts a touch panel provided in the character input device is referred to as “pen down”. Furthermore, when the handwriting input operation of the finger or the stylus pen is completed, the user's operation of leaving the finger or the stylus pen from the touch panel provided in the character input device is referred to as “pen-up”.

  In the following description, an array representing the order of at least one character such as full-width or half-width English letters, numbers, symbols, hiragana, katakana, or kanji is referred to as a “character string”.

  In the following description, the character string having the highest character recognition rate among the character recognition results for one character string of handwriting handwritten by a handwriting input operation with a finger or a stylus pen is referred to as “first candidate character string”. " Therefore, the first candidate character string is, for example, a character string in which a handwriting for one character handwritten in a handwriting input operation is specified by character recognition.

(Configuration of the character input device 1 of the present embodiment)
FIG. 1 is a block diagram showing the internal configuration of the character input device 1 of the present embodiment. The character input device 1 includes a coordinate input device 11, a coordinate input processing unit 12, a handwriting display processing unit 13, a character recognition processing unit 14, a recognized character display processing unit 15, a recognition candidate DB (Database) 15a, and a prediction conversion candidate generation unit 16. , A dictionary DB 16a, a predictive conversion candidate display processing unit 17, a context correction processing unit 18, a recognized character replacement determination unit 19, a recognized character correction unit 20, a first candidate display processing unit 21, and a display device 22.

  Among the character input device 1 of the present embodiment, the coordinate input processing unit 12, the handwriting display processing unit 13, the character recognition processing unit 14, the recognized character display processing unit 15, the prediction conversion candidate generation unit 16, and the prediction conversion candidate display processing unit 17 Each of the context correction processing unit 18, the recognized character replacement determining unit 19, the recognized character correcting unit 20, and the first candidate display processing unit 21 includes a program (not shown) built in the character input device 1 according to the program of the present invention. It can be executed by reading.

  The processor is, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or a DSP (Digital Signal Processor). A program according to the present invention, that is, a coordinate input processing unit 12, a handwriting display processing unit 13, a character recognition processing unit 14, a recognized character display processing unit 15, a prediction conversion candidate generation unit 16, a prediction conversion candidate display processing unit 17, a context correction A program for enabling each of the processing unit 18, the recognized character replacement determining unit 19, the recognized character correcting unit 20, and the first candidate display processing unit 21 is a ROM (Read Only) (not shown) built in the character input device 1. Memory).

  The coordinate input device 11 is provided corresponding to the screen (display screen) of the display device 22, and is configured using a touch panel that accepts a handwriting input operation by a user's finger FG (see FIG. 3 or 4) or a stylus pen. Yes. The touch panel is configured by, for example, an electrostatic capacity method capable of detecting a contact position on the screen of the finger FG or the stylus pen, but may be configured by using another method. The coordinate input device 11 detects a handwriting handwritten by a handwriting input operation with a finger FG or a stylus pen as coordinate information, and outputs it to the coordinate input processing unit 12.

  The coordinate input processing unit 12 includes a timer function of the character input device 1, controls the operation of each unit of the character input device 1, and temporarily stores coordinate information output from the coordinate input device 11 in a RAM (Random Access Memory) 12a. In addition, various input events including pen-down as an input event detected when the handwriting input operation is started or resumed or pen-up as an input event detected when the handwriting input operation is temporarily interrupted or terminated Perform appropriate processing. For example, the coordinate input processing unit 12 sends the coordinate information output from the coordinate input device 11 and information indicating that the handwriting input operation handwriting corresponding to the coordinate information is displayed on the display device 22 to the handwriting display processing unit 13. Output.

  In addition, the coordinate input processing unit 12 designates the first candidate character string 38 (see FIG. 2B) displayed in, for example, a pop-up display form based on the coordinate information output from the coordinate input device 11. It is determined whether or not it has been selected by the FG or the stylus pen. When the coordinate input processing unit 12 determines that the first candidate character string 38 displayed in the pop-up display form is selected by the finger FG or the stylus pen, the coordinate input processing unit 12 sends information to that effect to the recognized character correction unit 20. Output.

  Further, the coordinate input processing unit 12 determines whether or not a predetermined time (for example, 1 second) has elapsed since the coordinate input device 11 detected a pen-up input event based on the coordinate information output from the coordinate input device 11. Determine whether. When the coordinate input processing unit 12 determines that a predetermined time has elapsed since the coordinate input device 11 detected a pen-up input event, the coordinate input processing unit 12 outputs information to that effect to the first candidate display processing unit 21.

  The RAM 12 a is provided as a work memory during the operation of the coordinate input processing unit 12, and temporarily stores a program or data required during the operation of the coordinate input processing unit 12. For example, the RAM 12a temporarily stores the coordinate information output from the coordinate input device 11.

  The handwriting display processing unit 13 displays, based on the information output from the coordinate input processing unit 12, image data (for example, a bitmap) of handwriting input operations corresponding to the coordinate information output from the coordinate input device 11. It is displayed in the handwriting display area 34 (see FIG. 2A) on the screen of the apparatus 22. FIG. 2A is a diagram illustrating display contents displayed on the screen of the display device. The vertical and horizontal sizes of the handwriting display area 34 are defined in advance in the operation of the handwriting display processing unit 13, and may be fixed values or variable values.

  The character recognition processing unit 14 recognizes the handwriting of the handwriting input operation using the coordinate information output from the coordinate input processing unit 12, and specifies the character code corresponding to the handwriting. Since the character recognition in the character recognition processing unit 14 is a known technique as shown in, for example, Patent Document 1, detailed description thereof is omitted. The character recognition processing unit 14 outputs information on the character string (character code) of the identified character recognition result to the recognized character display processing unit 15, the predictive conversion candidate generation unit 16, and the context correction processing unit 18, respectively.

  Based on the information output from the character recognition processing unit 14 or the context correction processing unit 18, the recognized character display processing unit 15 performs context correction on the character recognition result character string corresponding to the handwriting input operation handwriting or the character string thereof. The text data of the character string is displayed in the recognized character display area 33 (see FIG. 2A) on the screen of the display device 22. The vertical and horizontal sizes of the recognized character display area 33 are defined in advance in the operation of the recognized character display processing unit 15, and may be fixed values or variable values.

  In addition, the recognized character display processing unit 15 touches when the information indicating that a touch operation or a tap operation has been performed on any of the character strings displayed in the recognized character display area 33 is received from the coordinate input processing unit 12. Information on a plurality of recognition candidates for the operated or tapped character string is acquired from the recognition candidate DB 15a, and a list of the plurality of recognition candidates is displayed as a recognition candidate list display area 37 on the screen of the display device 22 (see FIG. 2B). ). FIG. 2B is a diagram illustrating display contents displayed on the screen of the display device 22. The vertical and horizontal sizes of the recognition candidate list display area 37 are defined in advance in the operation of the recognized character display processing unit 15, and may be fixed values or variable values. In addition, in order to avoid complication of FIG. 1, illustration of the arrow between the coordinate input process part 12 and the recognition character display process part 15 is abbreviate | omitted.

  The recognized character display processing unit 15 receives information from the recognized character correcting unit 20 that the character string of the context correction result displayed in the recognized character display area 33 has been corrected to the first candidate character string 38. In this case, the character string of the context correction result displayed in the recognized character display area 33 is corrected and displayed as the first candidate character string 38.

  The recognition candidate DB 15a stores, for each character string, character string candidate data that has been subjected to character recognition in order of increasing character recognition rate. For example, in FIG. 2B, “3” (half-width numbers), “ro” (full-width hiragana) as character string candidates of the character recognition result for the character string “ro” displayed in the recognition candidate list display area 37. , “Tsu” (full-width hiragana), “tsu” (full-width hiragana), “fu” (full-width katakana) and “a” (full-width katakana) are shown.

  The predictive conversion candidate generation unit 16 refers to the dictionary DB 16a based on the information output from the character recognition processing unit 14 or the context correction processing unit 18, and the character string of the character recognition result or the character string whose context has been corrected Character string data of at least one predictive conversion candidate that matches forward with the column is extracted and generated.

  Further, the predictive conversion candidate generation unit 16 receives information from the recognized character correction unit 20 that the character string of the context correction result displayed in the recognized character display area 33 has been corrected to the first candidate character string 38. In this case, the dictionary DB 16a is referred to, and character string data of at least one predictive conversion candidate including the corrected character string (first candidate character string) is extracted and generated. The prediction conversion candidate generation unit 16 outputs the generated character string data of at least one prediction conversion candidate to the prediction conversion candidate display processing unit 17.

  The dictionary DB 16a is configured by using a hard disk or a flash memory built in the character input device 1, and stores a predictive conversion candidate character string that matches forward with a character string serving as a search key. For example, in the case where the search key is “3”, the dictionary DB 16a determines the character of the predictive conversion candidate (for example, “3 days”, “3”, “March”, etc.) that matches the search key “3”. Stores column data.

  The predictive conversion candidate display processing unit 17 displays the character string data output from the predictive conversion candidate generation unit 16 in the predictive conversion candidate display area 32 (see FIG. 2A) of the screen of the display device 22.

  When the context correction processing unit 18 receives from the character recognition processing unit 14 information indicating that the handwriting input by handwriting is two or more characters, based on the context with the character strings before and after the handwritten input handwriting, One of the character strings of the character recognition result in the character recognition processing unit 14 is corrected (context correction). The context correction in the context correction processing unit 18 is a well-known technique as shown in, for example, Patent Document 1, and thus detailed description thereof is omitted. In addition, the context correction | amendment process part 18 will receive the 1st character string by which the character recognition was carried out, when the information that the handwriting input by handwriting is the 1st character string from the character recognition process part 14 was received. No context correction.

  The context correction processing unit 18 outputs the character string data of the character string of the context correction result to the recognized character display processing unit 15 and the predictive conversion candidate generation unit 16, respectively, and further performs the context correction result for the handwritten input operation. Is output to the recognized character replacement determination unit 19.

  Here, an example of the operation of the context correction processing unit 18 will be briefly described with reference to FIGS. When the first handwritten handwritten input is the character string “3”, the recognized character display processing unit 15 displays the character string “3” as the character recognition result from the character recognition processing unit 14 on the display device 22. Is displayed in the recognized character display area 33 of the screen (see FIG. 2A).

  When the second handwritten handwritten input is the character string “B”, the context correction processing unit 18 reads the character string (for example, “B”) of the character recognition result in the character recognition processing unit 14 and before and after Based on the context with the character string “3” (immediately in FIGS. 2A and 2B), the character string “3” of the character recognition result in the character recognition processing unit 14 is corrected to the character string “RO”. (Context correction). The recognized character display processing unit 15 displays the character string “RO” of the context correction result in the context correction processing unit 18 so as to replace the character string “3” in the recognized character display area 33 of the screen of the display device 22 ( (See FIG. 2B). The operation of the context correction processing unit 18 will be described later with reference to FIGS.

  Based on the information output from the context correction processing unit 18, the recognized character replacement determination unit 19 executes the context correction on the handwritten input operation, and thereby the character recognition result of the character recognition processing unit 14. It is determined whether any character string has been replaced with another character string. The recognized character replacement determination unit 19 outputs to the first candidate display processing unit 21 determination result information indicating that any character string in the character recognition result in the character recognition processing unit 14 has been replaced with another character string.

  The recognized character correcting unit 20 is information output from the coordinate input processing unit 12, that is, information indicating that the first candidate character string 38 displayed in the pop-up display form is selected by the finger FG or the stylus pen. Based on the above, the character string of the character recognition result that is the target for which the first candidate character string 38 is displayed in the recognized character display area 33 is corrected to the first candidate character string 38. The recognized character correcting unit 20 outputs information indicating that the character string has been corrected to the first candidate character string 38 to the recognized character display processing unit 15, the predicted conversion candidate generation unit 16, and the first candidate display processing unit 21, respectively.

  The first candidate display processing unit 21 performs context correction processing based on the information output from the recognized character replacement determination unit 19 and the information output from the coordinate input processing unit 12 (information indicating that a predetermined time has elapsed since the pen-up). A first candidate character string 38 that is a character recognition result character string before context correction is displayed in a pop-up display form around the character string subjected to context correction in the unit 18. For example, the first candidate display processing unit 21 displays the first candidate character string 38 in a pop-up display form above the display position of the character string subjected to context correction. However, the display position of the first candidate character string 38 is not limited to the upper part of the display position of the character string subjected to context correction, and may be another display position. Details will be described later with reference to FIG.

  Further, the first candidate display processing unit 21 is based on the information output from the recognized character correcting unit 20, that is, information indicating that the character string of the character recognition result subjected to context correction has been corrected to the first candidate character string 38. In addition, the first candidate character string 38 displayed in the pop-up display form is not displayed.

  The display device 22 is configured using, for example, an LCD (Liquid Crystal Display) or an organic EL (Electroluminescence) display. The handwriting display processing unit 13, the recognized character display processing unit 15, the predictive conversion candidate display processing unit 17, and the first candidate display. In response to an instruction from at least one display processing unit in the processing unit 21, the display content shown in FIG. 2A or 2B is displayed on the screen.

  The screen of the display device 22 shown in FIG. 2A includes a confirmed character display area 31, a predicted conversion candidate display area 32 (see dotted line), a recognized character display area 33 (see alternate long and short dash line), and a handwriting display area 34 (two dots). And a key display area 35 are arranged. In the prediction conversion candidate display area 32, for example, in FIG. 2A, character strings of prediction conversion candidates are displayed over two stages, but may be displayed over one stage or three or more stages.

  In addition, on the screen of the display device 22 shown in FIG. 2B, when the character string “RO” is touched among the character strings “ROBAME” displayed in the recognized character display area 33, the character is displayed. A list of character recognition candidates for the column “RO” is displayed in the recognition candidate list display area 37 by the recognized character display processing unit 15. Further, on the screen of the display device 22 shown in FIG. 2B, the first candidate character string 38 of the character string “RO” among the character strings “ROBAME” displayed in the recognized character display area 33 is displayed. It is displayed in a pop-up display form. In FIG. 2B, the display position of the first candidate character string 38 is the upper part (directly above) of the character string “RO”.

(Overview of character input device operation)
Next, an outline of the operation of the character input device 1 of the present embodiment will be described with reference to FIGS. FIGS. 3A to 3C and FIGS. 4A to 4C are screen transition diagrams for explaining the operation outline of the character input device 1 of this embodiment in time series. Here, the operation outline of the character input device 1 is shown in the order of FIG. 3 (A), FIG. 3 (B), FIG. 3 (C), FIG. 4 (A), FIG. 4 (B), and FIG. The instruction medium for handwriting input operation is, for example, the user's finger FG.

  In FIG. 3A, the character string “3” is input by hand with the finger FG as the first handwriting of the handwriting input operation. The handwriting display processing unit 13 displays the image data of the character string “3” in the handwriting display region 34 based on the coordinate information from the coordinate input processing unit 12. The recognized character display processing unit 15 displays the character string “3” of the character recognition result in the character recognition processing unit 14 in the recognized character display area 33. The prediction conversion candidate display processing unit 17 outputs at least one character string (“3 days”, “3”, “March”, “3” in FIG. 3A) output from the prediction conversion candidate generation unit 16. ”,“ 31 days ”,“ 3D ”, etc.) are displayed in the prediction conversion candidate display area 32.

  In FIG. 3B, following the handwriting of the character string “3” shown in FIG. 3A, the character string “BA” is input by hand with the finger FG as the second handwriting of the handwriting input operation. . The handwriting display processing unit 13 displays the image data of the character string “BA” in the handwriting display region 34 based on the coordinate information from the coordinate input processing unit 12. The character recognition processing unit 14 specifies the character string “BA” of the second handwriting as the character recognition result. Based on the context of the character string “3” of the first handwriting and the character string “BA” specified by the character recognition processing unit 14 as the character string of the second handwriting, the context correction processing unit 18 The character string “3” of the first handwriting is corrected (context correction) to the character string “RO”. The recognized character display processing unit 15 displays the character string “3” displayed in the recognized character display area 33 based on the context correction result of the context correction processing unit 18 so as to replace the character string “RO”. .

  The recognized character replacement determination unit 19 receives information that the context correction processing unit 18 has corrected the character string “3” of the first handwriting to the character string “RO”, and receives the character recognition in the character recognition processing unit 14. It is determined that any character string of the result is replaced with another character string, and the determination result information is output to the first candidate display processing unit 21. The first candidate display processing unit 21 performs context correction when a predetermined time has elapsed since a pen-up input event was detected based on the information output from the recognized character replacement determination unit 19 and the coordinate input processing unit 12. The first candidate character string 38 (character string “3”), which is a character recognition result character string before context correction for the character string “RO”, is placed around the character string “RO” (for example, directly above). Displayed in a pop-up display form. The predictive conversion candidate display processing unit 17 outputs at least one character string output from the predictive conversion candidate generation unit 16 (in FIG. 3B, “donkey”, “donkey”, half-width “donkey”, “road” ”) Is displayed in the predictive conversion candidate display area 32.

  In FIG. 3C, following the handwriting of the character string “BA” shown in FIG. 3B, the character string “n” is input by hand with the finger FG as the third handwriting of the handwriting input operation. . Although not shown in FIG. 3C, the handwriting input operation for the third handwritten character string “n” is performed after the handwriting input operation for the second handwritten character string “BA” is completed. When started, the first candidate display processing unit 21 hides the first candidate character string 38 (character string “3”) for the character string “RO” displayed in the pop-up display form.

  When a predetermined time has elapsed since the handwriting input operation was completed with the finger FG and the pen-up input event was detected as the third handwriting, the character string “n” is similar to FIG. The first candidate display processing unit 21 surrounds (for example, directly above) the character string “RO” subjected to context correction, the first candidate that is a character string of the character recognition result before context correction for the character string “RO”. The character string 38 (character string “3”) is displayed in a pop-up display form. The prediction conversion candidate display processing unit 17 outputs at least one character string output from the prediction conversion candidate generation unit 16 (in FIG. 3C, “roadbed”, “roadside”, “abacus”, “roban”). The character string data of “Haban” in half-width display is displayed in the prediction conversion candidate display area 32.

  In FIG. 4A, following the handwriting of the character string “n” shown in FIG. 3C, the character string “Me” is input by hand with the finger FG as the fourth handwriting of the handwriting input operation. . As shown in FIG. 4A, when the handwriting input operation of the fourth handwritten character string “Me” is started, the first candidate display processing unit 21 is displayed in a pop-up display form. The first candidate character string 38 (character string “3”) for the character string “RO” is not displayed. The predictive conversion candidate display processing unit 17 outputs at least one character string output from the predictive conversion candidate generation unit 16 (in FIG. 4A, “road base”, “road base”, “robanme”, The character string data of “Robanme” and “Lobanme” in half-width display are displayed in the prediction conversion candidate display area 32.

  FIG. 4B shows a state in which the handwriting input operation with the finger FG of the character string “M” of the fourth handwriting shown in FIG. When the predetermined time has elapsed since the handwriting input operation was completed with the finger FG and the pen-up input event was detected as the fourth handwriting, the same as in FIG. The first candidate display processing unit 21 surrounds (for example, directly above) the character string “RO” subjected to context correction, the first candidate that is a character string of the character recognition result before context correction for the character string “RO”. The character string 38 (character string “3”) is displayed in a pop-up display form.

  Here, it is assumed that the finger FG performs a touch operation or a tap operation on the first candidate character string 38 (character string “3”) displayed in a pop-up display form. In response to the touch operation or the tap operation, as shown in FIG. 4C, the recognized character correction unit 20 displays the first candidate character string 38 (character string “3”) in the recognized character display area 33. The character string “RO” of the target character recognition result is corrected to the first candidate character string 38 (character string “3”). The first candidate display processing unit 21 hides the first candidate character string 38 (character string “3”) displayed in the pop-up display form in response to the touch operation or the tap operation.

  Further, the recognized character display processing unit 15 indicates that the context-corrected character string “RO” displayed in the recognized character display area 33 has been corrected to the first candidate character string 38 (character string “3”). When the information is received from the recognized character correcting unit 20, the character string of the context correction result displayed in the recognized character display area 33 is corrected and displayed as the first candidate character string 38. Further, the predictive conversion candidate generation unit 16 has received information from the recognized character correction unit 20 that the context-corrected character string displayed in the recognized character display area 33 has been corrected to the first candidate character string 38. In this case, the dictionary DB 16a is referred to, and character string data of at least one predictive conversion candidate including the first candidate character string 38 (character string “3”) is extracted and generated.

  The prediction conversion candidate display processing unit 17 displays the character strings (“road base”, “road surface”, “robanme”, “robanme”, and “half-width” displayed in the prediction conversion candidate display area 32. ) Is updated in the predictive conversion candidate display area 32 so as to be updated to the character string data of the character strings (“third”, “third”, “third”) output from the predictive conversion candidate generation unit 16. Let

(Operation procedure of the character input device 1 of this embodiment)
Next, the operation procedure of the character input device 1 of the present embodiment will be described with reference to FIGS. 5 and 6 are flowcharts for explaining the operation procedure of the character input device 1 of this embodiment. In the description of FIGS. 5 and 6, reference is made to FIGS. 3 and 4 as necessary.

  First, in FIG. 5, the character input device 1 monitors the presence or absence of an input event (for example, pen down) from the finger FG or the stylus pen in the coordinate input device 11, and waits until an input event is detected (S11). Only when the finger FG or the stylus pen detects an input event (for example, pen down) with respect to the handwriting display area 34 (S11, YES), the coordinate input device 11 receives the coordinate information corresponding to the input event as the coordinate input processing unit 12. Output to. The coordinate input processing unit 12 temporarily stores handwriting coordinate information corresponding to the input event (for example, pen down) in the RAM 12a (S12). The coordinate input processing unit 12 outputs the coordinate information output from the coordinate input device 11 and information indicating that the handwriting of the handwriting input operation corresponding to the coordinate information is displayed on the display device 22 to the handwriting display processing unit 13. .

  The handwriting display processing unit 13, based on each information output from the coordinate input processing unit 12, writes handwritten image data (for example, a bitmap) of a handwriting input operation corresponding to the coordinate information output from the coordinate input device 11. It is displayed in the handwriting display area 34 (see FIG. 3A).

  The coordinate input processing unit 12 determines, based on the coordinate information output from the coordinate input device 11, whether the finger FG or the stylus pen has detected a pen-up input event indicating a temporary interruption or termination of the handwriting input operation. (S14). When the coordinate input processing unit 12 does not detect a pen-up input event (S14, NO), the operation of the character input device 1 returns to step S1.

  On the other hand, when the coordinate input processing unit 12 detects a pen-up input event (S14, YES), the character recognition processing unit 14 uses the coordinate information output from the coordinate input processing unit 12 to perform a handwriting input operation. Is recognized, and a character code corresponding to the handwriting is specified (S15). The character recognition processing unit 14 determines whether or not the character string that has been character-recognized is the first character string of the handwriting input operation handwriting (S16).

  The character recognition processing unit 14 determines that the character string whose character is recognized is not the first character string of the handwriting input operation handwriting, that is, the character recognized character string is the second or later character string of the handwriting input operation handwriting. (S16, NO), the character string (character code) information of the identified character recognition result is sent to the recognized character display processing unit 15, the predictive conversion candidate generation unit 16, and the context correction processing unit 18. Output.

  When the context correction processing unit 18 receives from the character recognition processing unit 14 information indicating that the handwriting input by handwriting is two or more characters, based on the context with the character strings before and after the handwritten input handwriting, One of the character strings of the character recognition result in the character recognition processing unit 14 is corrected (context correction) (S17). The context correction processing unit 18 outputs the character string data of the character string of the context correction result to the recognized character display processing unit 15 and the predictive conversion candidate generation unit 16, respectively, and further performs the context correction result for the handwritten input operation. Is output to the recognized character replacement determination unit 19.

  On the other hand, when the character recognition processing unit 14 determines that the character string that has been character-recognized is the first character string of the handwriting input operation handwriting (S16, YES), the determination result is context corrected. The data is output to the processing unit 18. In this case, the context correction processing unit 18 does not context-correct the first character string that has been character-recognized.

  After step S <b> 17, the recognized character display processing unit 15, based on the information output from the character recognition processing unit 14 or the context correction processing unit 18, the character recognition result character string corresponding to the handwriting input operation handwriting or the character thereof The text data of the character string whose column has been context corrected is displayed in the recognized character display area 33 (see FIG. 3A).

  Further, the predictive conversion candidate generation unit 16 refers to the dictionary DB 16a based on the information output from the character recognition processing unit 14 or the context correction processing unit 18, and the character string of the character recognition result or its character string is context corrected. The character string data of at least one predictive conversion candidate that matches the forward character string is extracted and generated. The prediction conversion candidate generation unit 16 outputs the generated character string data of at least one prediction conversion candidate to the prediction conversion candidate display processing unit 17. The predictive conversion candidate display processing unit 17 displays the character string data output from the predictive conversion candidate generation unit 16 in the predictive conversion candidate display area 32 (see FIG. 3A) of the screen of the display device 22 (S18). .

  Next, in FIG. 6, the recognized character replacement determination unit 19 performs the character recognition process by executing the context correction on the handwritten input operation based on the information output from the context correction processing unit 18. It is determined whether any character string of the character recognition result in the unit 14 is replaced with another character string (S19). Only when the recognized character replacement determination unit 19 determines that any character string of the character recognition result in the character recognition processing unit 14 has been replaced with another character string (S19, YES), the character recognition in the character recognition processing unit 14 The determination result information indicating that any character string of the recognition result is replaced with another character string is output to the first candidate display processing unit 21.

  The first candidate display processing unit 21 is based on the information output from the recognized character replacement determination unit 19 and the information output from the coordinate input processing unit 12 (information indicating that a predetermined time has elapsed since pen-up) (YES in S20). ), A first candidate character string 38 that is a character recognition result character string before context correction is displayed in a pop-up display form around the character string subjected to context correction in the context correction processing unit 18. (S21).

  During the predetermined time of step S20, when the user continuously inputs characters by handwriting input operation, the first candidate character string 38 (character string “3”) is displayed in a pop-up display form during the handwriting input operation. This delay time is provided so that the user does not feel bothered by continuing to do so.

  Thereby, the character input device 1 does not continue to display the first candidate character string during the pen-down when the character is continuously handwritten during the user's handwriting input operation, and the predetermined time elapses after the pen-up. Then, since the first candidate character string is displayed, the handwriting input operation can be performed without causing the user to feel bothered by the display of the first candidate character string.

  When the coordinate input processing unit 12 detects a pen-down input event, that is, a handwriting input operation for the next character, based on information output from the coordinate input device 11 before the predetermined time has elapsed in step S20. (S21, YES), the timer timing is forcibly stopped. Thereafter, the operation of the character input device 1 returns to step S12.

  After step S22, the coordinate input processing unit 12 uses the finger FG or the stylus pen to display the first candidate character string 38 displayed in a pop-up display form based on the coordinate information output from the coordinate input device 11, for example. It is determined whether or not it has been selected (S23). When it is determined that the character string 38 of the first candidate (see FIG. 2B) displayed in the pop-up is not selected by the finger FG and the stylus pen, the operation of the character input device 1 ends.

  When the coordinate input processing unit 12 determines that the first candidate character string 38 displayed in the pop-up has been selected by the finger FG or the stylus pen (S23, YES), the information to that effect is sent to the recognized character correction unit 20. Output. The recognized character correcting unit 20 is information output from the coordinate input processing unit 12, that is, information indicating that the first candidate character string 38 displayed in the pop-up display form has been selected by the finger FG or the stylus pen. Based on the above, the character string of the character recognition result that is the target for which the first candidate character string 38 is displayed in the recognized character display area 33 is corrected to the first candidate character string 38 (S24). The recognized character correcting unit 20 outputs information indicating that the character string has been corrected to the first candidate character string 38 to the recognized character display processing unit 15, the predicted conversion candidate generation unit 16, and the first candidate display processing unit 21, respectively.

  The recognized character display processing unit 15 receives information from the recognized character correcting unit 20 that the corrected character string displayed in the recognized character display area 33 has been corrected to the first candidate character string 38. Then, the character string after context correction displayed in the recognized character display area 33 is corrected and displayed as the first candidate character string 38 (S25).

  The first candidate display processing unit 21 corrects the information output from the recognized character correction unit 20, that is, the character string on which the first candidate character string 38 is displayed, to the first candidate character string 38. Based on the information to the effect, the first candidate character string 38 displayed in the pop-up display form is hidden (S26). Thereby, the character input device 1 does not attract the user's attention more than necessary by hiding the first candidate character string 38 that is no longer required to be displayed, thereby simplifying the user's handwriting input operation. Can be made.

  The predictive conversion candidate generation unit 16 receives information from the recognized character correction unit 20 that the character string after context correction displayed in the recognized character display area 33 has been corrected to the first candidate character string 38. Then, by referring to the dictionary DB 16a, at least one prediction conversion candidate character string data including the corrected character string (first candidate character string) is extracted and generated. The prediction conversion candidate generation unit 16 outputs the generated character string data of at least one prediction conversion candidate to the prediction conversion candidate display processing unit 17.

  The predictive conversion candidate display processing unit 17 displays the character string data output from the predictive conversion candidate generation unit 16 in the predictive conversion candidate display area 32 of the screen. For example, the prediction conversion candidate display processing unit 17 displays the character strings (“road base”, “road base”, “robanme”, “robanme”, “half-width display” displayed in the prediction conversion candidate display area 32. Rovanme ") is updated to the character string data of the character strings (" third "," third "," third ") output from the predictive conversion candidate generation unit 16. (S27). Thereby, the character input device 1 corrects the recognized character and further updates the predicted conversion candidate when the user selects the first candidate character string by the touch operation or the tap operation. Selection can be simplified, the labor of correction can be reduced, and the correction work can be performed easily and in a short time.

  As described above, the character input device 1 according to the present embodiment determines that context correction has been performed on the input character string, and before the context correction, around the display position of the character string of the character recognition result subjected to context correction. Is displayed as a first candidate character string. Therefore, the character input device 1 allows the user to easily recognize that the context correction for the input character string has been executed, and if necessary, prompts the user to make a correction instruction for the character string after the context correction. Can be encouraged.

  Although various embodiments have been described with reference to the drawings, it goes without saying that the present invention is not limited to such examples. It is obvious for those skilled in the art that variations and modifications of various embodiments, and combinations of various embodiments can be conceived within the scope of the claims. Of course, it is understood that it belongs to the technical scope of the present invention.

  For example, in the above-described embodiment, it has been described that the display position of the first candidate character string 38 is the upper part (directly above) of the recognized character displayed in the recognized character display area 33. If it is, it will not be specifically limited. FIG. 7 is a diagram illustrating variations in the display position of the first candidate character string. FIG. 7A shows an example displayed at the top of the recognized character string “RO”. FIG. 7B is an example displayed below the recognized character string “RO”. FIG. 7C is an example displayed on the right side of the recognized character string “RO”. FIG. 7D is an example displayed in the upper right part of the recognized character string “RO”.

  In FIG. 7A, the first candidate display processing unit 21 displays the first candidate character string 38 (character string “3”) above the character string after context correction displayed in the recognized character display area 33. However, the prediction conversion candidate display area 32 is displayed so as not to overlap. Thereby, the character input device 1 can prevent the first candidate character string 38 from blocking the display of at least one prediction conversion candidate displayed in the prediction conversion candidate display area 32.

  Although not shown in FIG. 7A, the first candidate display processing unit 21 recognizes the recognized character display area even when the prediction conversion candidate display area 32 and the display area of the first candidate character string 38 overlap. In the case of displaying on the upper part of the character string after context correction displayed in 33, the first candidate character string 38 (character string “3”) is displayed in a semi-transparent (translucent) display form. Also good. Accordingly, the character input device 1 can similarly prevent the first candidate character string 38 from blocking the display of at least one prediction conversion candidate displayed in the prediction conversion candidate display area 32.

  In FIG. 7B, the first candidate display processing unit 21 converts the first candidate character string 38 (character string “3”) into the character recognition result after context correction displayed in the recognized character display area 33. The lower part of the character string is displayed so as not to overlap the prediction conversion candidate display area 32. Thereby, the character input device 1 can prevent the first candidate character string 38 from blocking the display of at least one prediction conversion candidate displayed in the prediction conversion candidate display area 32.

  In FIG. 7C, the first candidate display processing unit 21 converts the first candidate character string 38 (character string “3”) into the character recognition result after context correction displayed in the recognized character display area 33. It is displayed on the right side of the character string and between the adjacent character strings on the right side. Although not shown in FIG. 7C, the recognized character display processing unit 15 displays the character string that is displayed in the recognized character display area 33 as the character string of the context correction result in which the first candidate character string is displayed. In the case of the last character string of the column (for example, the character string “M” in FIG. 7C), the character string of the character recognition result is scrolled to the left and displayed as necessary. Thereby, the character input device 1 can prevent the first candidate character string 38 from blocking the display of at least one prediction conversion candidate displayed in the prediction conversion candidate display area 32.

  In FIG. 7D, as in FIG. 7A, the first candidate display processing unit 21 displays the first candidate character string 38 (character string “3”) in the recognized character display area 33. The character string after context correction is displayed in the upper right part so as not to overlap the prediction conversion candidate display area 32. Thereby, the character input device 1 can prevent the first candidate character string 38 from blocking the display of at least one prediction conversion candidate displayed in the prediction conversion candidate display area 32.

  FIG. 8 is a diagram illustrating an example in which a plurality of first candidate character strings 39 are displayed above a plurality of recognized character strings. In FIG. 8, when the user inputs the character string “3” four times in succession by the handwriting input operation, the character recognition processing unit 14 converts the first character string of the handwriting input operation handwriting to the character string “3”. ”And then the second to fourth character strings are similarly specified as the character string“ 3 ”. However, the context correction processing unit 18 context-corrects each of the second to fourth character strings “3” to the character string “RO” as in the above-described embodiment.

  In this case, as illustrated on the right side of the arrow in FIG. 8, the first candidate display processing unit 21 applies a plurality of first candidates to the second to fourth character strings “RO” whose context has been corrected. The character string 39 (character string “3”) is displayed. In FIG. 8, the first candidate display processing unit 21 is not limited to the case where the second to fourth character strings are all the same, and even if the character strings are different, One candidate character string may be displayed.

  In the above-described embodiment, the timing of displaying the first candidate character around the recognized character is after a predetermined time since the replacement with the recognized character. For example, the character string of the character recognition result is a recognized character. While displayed in the display area 33, it may always be displayed. Thereby, the character input device 1 can easily attract the user's attention to the first candidate character string, and can easily urge correction.

  Further, in the above-described embodiment, the timing at which the first candidate character string is not displayed is the time when the recognized character correcting unit 20 corrects the character string whose context has been corrected to the first candidate character string and the recognized character display processing unit 15. Was when the character recognition result string was updated and displayed. However, the timing at which the first candidate character string is not displayed is, for example, when the coordinate input device 11 detects a pen-down input event, the coordinate input device 11 determines the input event of the confirming operation to confirm the handwritten character string. When detected, the coordinate input device 11 is displayed in the prediction conversion candidate display area 32 when the coordinate input device 11 detects an input event of the deletion operation of the character string of the context correction result in which the first candidate character string is displayed. For example, when a predictive conversion candidate selection operation is detected, a character input application is not displayed on the character input device 1. Further, the timing for hiding the first candidate character string may be when an arbitrary character string candidate is selected from the recognition candidate list display shown in FIG. Thereby, the character input device 1 can hide the display of a 1st candidate at the appropriate timing according to each case, and can improve a user's operativity.

  In the above-described embodiment, the context correction processing unit 18 determines the context between the first character string and the second character string after the second and subsequent character strings are recognized by handwriting input operation. Based on this, the first character string was context corrected. However, the context correction processing unit 18, for example, after the third character string is recognized, based on the context with the first and second character strings before the third character string. The third character string may be context corrected.

  In the embodiment described above, the display form of the first candidate character string 38 is not limited to the pop-up, and the first candidate display processing unit 21 adds a specific mark (for example, a square or a circle) to the first candidate. One candidate character string may be displayed or blinked. The first candidate display processing unit 21 reverse-displays the recognized character replaced by the recognized character correcting unit 20, displays it with a specific color, or changes the recognized character or the surrounding background. Thus, the user can easily recognize that the character string recognized by the context correction is replaced with another character string.

  In the above-described embodiment, the case where characters such as hiragana and kanji are input has been described. However, the present invention can be similarly applied when inputting special characters such as graphics and symbols.

  The present invention, when accepting handwritten input, easily recognizes that context correction has been performed on the input character string, and if necessary, a character recognition device that prompts a correction instruction for the character string after context correction, It is useful as a character recognition method and program.

1 Character Input Device 12 Coordinate Input Processing Unit 12a RAM
13 Handwriting display processing unit 14 Character recognition processing unit 15 Recognized character display processing unit 15a Recognition candidate DB
16 Predictive conversion candidate extraction unit 16a Dictionary DB
17 Predictive Conversion Candidate Display Processing Unit 18 Context Correction Processing Unit 19 Recognized Character Replacement Determination Unit 20 Recognized Character Correction Unit 21 First Candidate Display Processing Unit 22 Display Device

Claims (14)

A display for displaying data on the screen;
A touch panel that is provided corresponding to the screen and receives handwritten input;
A character recognition processing unit for recognizing the handwritten handwritten character;
A recognized character display processing unit for displaying the character recognition result of the character recognition on the screen;
When the handwritten input handwriting is two or more characters, based on the context with the character string before and after the handwriting, a context correction processing unit that corrects any character string in the handwriting,
A replacement determination unit that determines whether any of the character strings has been context-corrected;
When it is determined that any one of the character strings has been context corrected, the character recognition result character string before the context correction is placed around the display position of the character recognition result character string after the context correction. A character recognition device comprising: a first candidate display processing unit for displaying as one candidate character string. The character recognition device according to claim 1,
A prediction conversion candidate generation unit that generates at least one prediction conversion candidate including the character recognition result;
A recognized character correcting unit that corrects the character string of the character recognition result subjected to the context correction to the first candidate character string in response to the selection operation of the displayed first candidate character string; Prepared,
The prediction conversion candidate generation unit updates the generated at least one prediction conversion candidate to a prediction conversion candidate including the first candidate character string, based on the correction to the character string of the first candidate. Character recognition device. The character recognition device according to claim 1 or 2,
The first candidate display processing unit is a character recognition device that displays the first candidate character string when a predetermined time has elapsed since the touch panel detected a temporary interruption or termination of the handwriting input. The character recognition device according to claim 1 or 2,
The first candidate display processing unit is a character recognition device that always displays the first candidate character string. The character recognition device according to claim 1 or 2,
When it is determined that a plurality of character strings in the handwriting are context-corrected, the first candidate display processing unit, for each of the plurality of character strings subjected to the context correction, the first corresponding to each character string A character recognition device that displays one candidate character string. The character recognition device according to any one of claims 3 to 5,
The first candidate display processing unit displays the first candidate character at a position above or below a display position of the character string of the character recognition result subjected to the context correction. The character recognition device according to any one of claims 3 to 5,
The first candidate display processing unit, when there is another character string that is continuous with the character string of the character recognition result subjected to the context correction in the handwritten input handwriting, the character subjected to the context correction A character recognition device for displaying the first candidate character string between a character string of a recognition result and the other character. The character recognition device according to claim 1 or 2,
The first candidate display processing unit is a character recognition device that hides the first candidate character string displayed on the screen when the touch panel detects resumption of the handwriting input. The character recognition device according to claim 1 or 2,
The first candidate display processing unit is configured to hide the first candidate character string displayed on the screen when the touch panel detects a confirmation operation or a deletion operation on the character string of the character recognition result. . The character recognition device according to claim 2,
The first candidate display processing unit is configured to display the first displayed on the screen when the touch panel detects a selection operation of one of the prediction conversion candidates generated by the prediction candidate generation unit. A character recognition device that hides candidate character strings. The character recognition device according to claim 2,
The first candidate display processing unit hides the first candidate character string displayed on the screen when the touch panel detects a selection operation of the first candidate character string displayed on the screen. Character recognition device. The character recognition device according to claim 2,
The recognized character display processing unit displays a list of character string candidates of the character recognition result on the screen,
The first candidate display processing unit is displayed on the screen when the touch panel detects a selection operation of any candidate from a list of character string candidates of the character recognition result displayed on the screen. And a character recognition device for hiding the first candidate character string. A character recognition method in a character recognition device having a display unit for displaying data on a screen,
A step corresponding to the screen and receiving a handwriting input;
Recognizing the handwritten handwritten character,
Displaying the character recognition result of the character recognition on the screen;
When the handwritten input handwriting is two or more characters, based on the context with the character string before and after the handwriting, correcting any one of the handwriting,
Determining whether any of the character strings has been context corrected;
When it is determined that any one of the character strings has been context corrected, the character recognition result character string before the context correction is placed around the display position of the character recognition result character string after the context correction. A character recognition method comprising: displaying as a candidate character string. In a character recognition device that is a computer having a display unit for displaying data on a screen,
A step corresponding to the screen and receiving a handwriting input;
Recognizing the handwritten handwritten character,
Displaying the character recognition result of the character recognition on the screen;
When the handwritten input handwriting is two or more characters, based on the context with the character string before and after the handwriting, correcting any one of the handwriting,
Determining whether any of the character strings has been context corrected;
When it is determined that any one of the character strings has been context corrected, the character recognition result character string before the context correction is placed around the display position of the character recognition result character string after the context correction. And a step of displaying as a candidate character string.

Images