JP3242525B2 - Character string presentation method for online recognition - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character in online recognition for recognizing a drawing trajectory input using an input device such as a pen / tablet as a character, figure, or character string and outputting the recognition result. It relates to a column presentation method.

[0002]

2. Description of the Related Art As an on-line character / graphic recognition method, for example, Japanese Patent Application No. 5-238419 discloses a method of recognizing not only handwriting input online as a character or a graphic but also a pre-registered character. A method of recognizing as a column is disclosed. According to this method, a character string composed of a plurality of registered characters can be output by inputting a handwriting for one character or one figure, and the operation burden on the input person is greatly reduced.

In this case, the handwriting used for recognizing a character string is a handwriting that can be recognized as a normal character or graphic (a character string can also correspond to a character or graphic code). It may have a shape that cannot be recognized as a normal character or figure (a handwriting can be directly associated with a character string itself). Alternatively, both of them may be mixed and registered.

[0004]

However, in the above-described character string recognition method, various methods have been proposed for registering a character string. However, the method of presenting a character string as a recognition result is the same as the conventional method. Is adopted. FIG. 8 is a schematic diagram showing an example of presenting a character string as a recognition result candidate by the conventional character string presentation method. Here, it is assumed that a pen / tablet is used as an input device, and an area (frame display) where a group of recognition result candidates is presented is provided on the tablet. (A) shows a presentation example of a recognition result candidate group when a character string is not included as a recognition result.
Each of the presentation examples of the recognition result candidate group when the character string is included as the recognition result is shown.

[0005] As shown in the figure, in (a), candidate characters as recognition results are "matsu",
They are displayed at predetermined display positions at predetermined intervals such as "forest", "cedar", "school", and "village". On the other hand, in (b), a seven-character character string “Matsushita Electric Industrial Co., Ltd.” is presented as the second recognition candidate. (B)
As shown by, when a character string is presented as a recognition result candidate mixed with a character or a figure, the recognition result candidate becomes difficult to see, and the number of candidates presented decreases.
This is a problem particularly when the presentation area cannot be made large as in a portable device. Therefore, complicated control is required as a candidate selection operation performed by the input person, which is inconvenient. For example, as shown in FIG.
Controls such as selecting the directions of the arrows displayed at the lower left and right ends of the presented candidate group with a pen, moving the presented content rightward or leftward, and displaying other unpresented candidates are displayed. Conceivable.

The present invention has been made in view of the above situation, and provides a character string presentation method which makes it easy to present a character string as a recognition result candidate in online recognition and can easily present it. The purpose is to:

[0007]

In order to achieve the above object, the present invention according to claim 1 uses an input device,
This is a character string presentation method in online recognition that recognizes an input drawing locus as a character, a figure, or a character string and outputs the recognition result, and whether a character string is included in a recognition result candidate group. A determining step of determining whether
When it is determined that the character string is included in the recognition result candidate group, a symbol indicating the character string is presented in the first presentation area together with another recognition result candidate character or graphic group. 1, a selecting step of selecting a candidate to be determined from the presented group of recognition result candidates, and when the symbol indicating a character string is selected, the corresponding character string is changed to a second character string. And a second presenting step of presenting the information in the presenting area.

According to a second aspect of the present invention, in the character string presenting method according to the first aspect, in the first presenting step, two or more character strings are included in the recognition result candidate group. When included, for each of the character strings, the symbol indicating the character string is presented in the first presentation area;
In the second presentation step, a character string corresponding to the selected symbol is presented in the second presentation area.

Further, in the invention according to claim 3, in the first presenting step, when two or more character strings are included in the recognition result candidate group, all of the character strings are represented. Then, one of the symbols indicating a character string is presented in the first presentation area, and in the second presentation step,
When the symbol indicating a character string is selected,
It is characterized in that all relevant character strings are presented in the second presentation area.

Further, the invention according to a fourth aspect of the present invention includes, following the determining step, a second determining step of determining whether or not the recognition result candidate group includes only a character string. When it is determined that the group is composed of only character strings, the method includes a third presentation step of presenting the corresponding character strings in the second presentation area.

[0011]

According to the configuration of the first aspect of the present invention, when the drawing trajectory input by using the input device is recognized as a character, a graphic, or a character string, the determination step includes:
A determination is made as to whether a character string is included in the recognition result candidate group. Therefore, when it is determined that the character string is included in the recognition result candidate group, in the first presentation step, the symbol indicating the character string is replaced with another recognition result candidate character or It is presented in the first presentation area together with the figure group. Subsequently, in a selection step, a candidate to be determined is selected from the group of recognition candidates presented in the first presentation area. Then, when a symbol indicating a character string is selected, the corresponding character string is further presented in the second presentation area in the second presentation step.

According to the configuration of the second aspect of the present invention, when two or more character strings are included in the recognition result candidate group, the character string is determined in the first presentation step. Are presented in correspondence with the individual character strings. Further, when any of the presented symbols is selected in the selection step, a character string corresponding to the selected symbol is presented in the second presentation area in the second presentation step.

According to the third aspect of the present invention, in the case where two or more character strings are included in the recognition result candidate group, the character string may be included in the first presentation step. Is presented in the first presentation area as a representative of all character strings. Further, when the presented symbol is selected in the selection step, all the corresponding character strings are presented in the second presentation area in the second presentation step.

Therefore, if not only characters or figures but also a registered character string are included in the recognition result candidate group with respect to the input handwriting, a symbol indicating the character string is displayed. Are presented together with a character or graphic group as another recognition result candidate. Therefore, when the input user selects the symbol, the content of the corresponding character string is separately presented. Therefore, the input user can easily see the details of the character string, and can easily select and confirm an expected character string from the details.

According to the fourth aspect of the present invention, following the determining step of the first aspect, the second step
Is determined, and it is determined whether or not the recognition result candidate group is composed of only character strings.
Then, when it is determined that the recognition result candidate group is composed only of the character strings, the corresponding character strings are directly presented in the second presentation area in the third presentation step.

As a result, the first presenting step of presenting the symbol indicating the character string is omitted, and the candidate character string is directly presented in the second presenting area. The character string can be determined quickly and easily by an easy operation.

[0017]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing a configuration of an online / character / graphic recognition apparatus to which a character string presentation method according to the present invention is applied. This device includes an input data extraction unit 1
, An online recognition unit 2, a recognition dictionary 3, a recognition result candidate presentation unit 4, and a selection unit 5. here,
The input data extraction unit 1 extracts input data representing the characteristics of handwriting input by handwriting online. Further, the online recognition unit 2 compares the input data extracted by the input data extraction unit 1 with the contents of the recognition dictionary 3 to extract a recognition result candidate group. The recognition dictionary 3 stores the characteristics of handwriting data to be associated with the characters, figures, or character strings to be determined. In addition, the recognition result candidate presentation unit 4 displays and presents the candidate group of the recognition result specified by the online recognition unit 2 to the input person. Then, the selection unit 5 allows the input person to select a candidate to be determined from the recognition result candidate group presented by the recognition result candidate presentation unit 4.

Next, the operation of the online character / graphic recognition apparatus having the above configuration will be described. There are various devices such as a mouse, a digitizer, and a trackball as online input devices. Here, a combination of a pen and a tablet is used. First, the input data extraction unit 1 extracts input data used by the online recognition unit 2 from the input handwriting data. in this case,
As input data, for example, a series of coordinate points from pen down to pen up (hereinafter referred to as “stroke”)
Can be used. In some cases, information such as the sampling time may be added to each coordinate point. Subsequently, the input data extraction unit 1 classifies the extracted stroke itself or a predetermined number of tens to hundreds of stroke types into strokes, and then uses the stroke type as input data as an input to the online recognition unit 2. Pass to

Therefore, the online recognition unit 2 stores the input data, the contents of the recognition dictionary 3 (characters and figures to be recognized, and the characteristics of handwriting data corresponding to a character string registered in advance). ) Are compared and extracted as a recognition result candidate group. As a recognition method in this case, for example, a method utilizing the shape information and position information of the stroke can be used (Tomimoto et al., “Online Handwritten Kanji / Hiragana Recognition”, IEICE Technical Report,
RL-82-37).

Further, the recognition target may include a registration target additionally registered by the input person. For example, when a certain character string is to be recognized, the character string can be registered using an ordinary character or graphic that has already been recognized. Specifically, if the character string “matsu” is directly recognized and the character strings “Matsushita Electric Industrial Co., Ltd.” and “Ichiro Matsushita” are registered in association with each other, the handwriting corresponding to the character “matsu” is registered. When data is entered, the character string "Matsushita Electric Industrial Co., Ltd."
And "Ichiro Matsushita" can be included in the recognition result candidate group.

Further, as shown in FIG. 2, a handwriting which is not recognized as a normal character or graphic can be registered directly in correspondence with a character string. FIG. 2 is an explanatory view showing a direct correspondence between a handwriting input by a pen and a character string, and an example of registration of the character string. As shown in the figure, the handwriting data of “Pine of character surrounded by a substantially circle” drawn on the tablet is not normally recognized as a character / graphic recognition target. However, if this input data is registered in correspondence with, for example, a character string of “Matsushita Electric Industrial Co., Ltd.”, by drawing “Pine of a substantially circled character” on a tablet, It is possible to display “Electrical Industry Co., Ltd.” as a recognition result.

On the other hand, if the on-line recognition unit 2 cannot find a recognition target that can be extracted as a recognition result candidate, it sets the output for the input data to “reject” and sets this as a (single) recognition result candidate. Extract and leave it to further processing. In this case, "Reject"
In this case, a special symbol indicating this is usually presented to prompt the input person to re-enter.

Next, the recognition result candidates extracted by the online recognition unit 2 are passed to the recognition result candidate presentation unit 4 and presented to the input person. Here, when the registered character string is included in the recognition result candidate group, the character string is not directly presented, and as shown in FIG. Is presented in place of the character string. FIG.
FIG. 4 is a schematic diagram showing a presentation example of a recognition result candidate group including a symbol indicating a character string. As shown in the figure, the recognition result candidates extracted by the online recognition unit 2 are extracted from the recognition result candidate presentation unit 4 in the order in which they are extracted. It is presented as a symbol indicating a column, "forest,""cedar,""school," and "village."

Therefore, the selection unit 5 determines and outputs the final recognition result as a final recognition result, the recognition result candidate selected by the input person using the pen from the presented recognition result candidate group. In this case, the selecting unit 5 selects not only the candidate directly selected by the input user using the pen but also the case where the input operation has timed out or the next input shift occurs, from the recognition result candidate group. One may be selected based on some criteria (for example, one that is first passed from the online recognition unit 2 to the recognition result candidate presentation unit 4) and output as the final recognition result.

On the other hand, when a symbol indicating the character string is selected by the input person from the group of recognition result candidates, the recognition result candidate presenting section 4 converts the specific contents of the corresponding character string into the following figure. Displayed separately as shown in FIG. FIG. 4 is a schematic diagram showing a presentation example of specific contents of a character string. As shown in the figure, when the input user selects a symbol indicating a character string using a pen, the registered "Matsushita Electric Industrial Co., Ltd."
Is successively presented separately. Here, an example is shown in which a registered character string is presented below the presentation column of the recognition result candidate group. When the input person selects the presented character string, the character string is determined as a recognition result.

When two or more registered character strings are included in the recognition result candidate group, a plurality of symbols indicating the character strings may be presented corresponding to each character string. In this case, the symbol on the left with a high recognition probability is selected first, and the corresponding character string is presented. If the expected character string is not included in the character string, another symbol is sequentially selected. The corresponding character string will be presented.

Furthermore, when two or more registered character strings are included in the recognition result candidate group, only one symbol indicating the character string is presented, thereby reducing the complexity of operation. It is possible to avoid this and make effective use of the presentation surface. That is, when the input user selects the symbol indicating the character string in the selection unit 5, the recognition result candidate presentation unit 4 displays all the corresponding character strings as shown in FIG. FIG.
It is a schematic diagram which shows the example of presentation of a some character string. As shown in the figure, when the input user selects a symbol indicating a character string using a pen, two registered character strings “Matsushita Electric Industrial Co., Ltd.” and “Ichiro Matsushita” are separately presented. Is done. Here, an example is shown in which a character string “Matsushita Electric Industrial Co., Ltd.” having a high recognition probability is presented in the upper part, and another character string “Ichiro Matsushita” is presented below. If there are many candidate character strings, a scrolling process is performed to deal with them.

Therefore, in the selecting section 5, it is assumed that the input person can select the final recognition result from all the presented character strings. In this way, the input user can list the contents of all the corresponding character strings, and the selection is reliable and easy. On the other hand, in a case where the recognition result candidate group extracted by the online recognition unit 2 is entirely composed of only character strings, even if only one symbol indicating the character string or the number of character strings is presented. In the end, the input user always selects the symbol indicating the character string presented by the selection unit 5. In such a case, the recognition result candidate presentation unit 4
Here, it is assumed that the presentation of the symbol indicating the character string is omitted, and the contents of one or more character strings are directly displayed as shown in FIG.

FIG. 6 is an explanatory diagram showing a direct correspondence between a handwriting input by a pen and a plurality of character strings, and an example of presentation of the character strings. As shown in the figure, by drawing "Pine of characters surrounded by a circle" on the tablet, all the character strings "Matsushita Electric Industrial Co., Ltd.""IchiroMatsushita" is recognized, and the recognition result candidate presentation unit 4 directly presents them. This eliminates the need for the input user to select an intervening trivial selection, that is, a “symbol indicating a character string”, thereby further improving convenience.

FIG. 7 is a flowchart showing control for presenting recognition result candidates by the character string presenting method according to the present invention. From the online recognition unit 2 shown in FIG.
In the recognition result candidate presentation unit 4 that has received the recognition result candidate,
First, it is determined whether or not a character string is included in the received recognition result candidate group (S1). In this determination, whether the candidate character is composed of one character based on the information received from the online recognition unit 2 or not.
Or, it is determined whether it is composed of two or more.
If a character string is included (Ye in S1)
In the case of s), it is further determined whether or not the candidate is composed only of a character string, that is, whether or not the candidate is composed entirely of a plurality of characters (S2).

Here, if the candidate is not composed only of a character string but is a candidate together with other characters and figures (No in S2), a symbol indicating a character string, for example, “A” is presented in the first presentation area together with other recognition candidate characters and graphics (S3). Subsequently, when the symbol “A” is selected by the input unit 5 in the selection unit 5 (Yes in S4), the corresponding character string is presented in the second presentation area (S5). Step S
If Yes in 2, that is, if it is determined that the candidate is composed of only a character string, the corresponding character string is immediately presented to the second presentation area (S5). Then, the process waits for the input user to select the next process (S6).

If No in step S1,
That is, when a character string is not included in the recognition result candidate group, a candidate character or graphic is presented in the first presentation area (S10), and the process proceeds to step S6. In the above embodiment, the character string presented as the recognition result candidate is a character string in which a plurality of characters or figures are arranged. There is no harm in including those that do not exist. In such a case, when the symbol indicating the character string is selected, the presentation is made as a single character or graphic, or merely as an “empty character string (or nothing is presented)”.

In the above embodiment, the recognition target other than ordinary characters or figures presented as recognition result candidates is represented by a symbol "character string". It goes without saying that any symbol that can be used may be used, and the output result may be a data sequence such as “graphic sequence”, “image sequence”, “audio data sequence”, or any of these. A column in which various data columns are mixed may be used. That is, any data can be easily selected by the input person by associating the data with a handwriting of a specific shape or a character or a graphic which has already been recognized. It becomes possible.

[0034]

According to the present invention described above, a symbol indicating a character string to be a recognition candidate is treated and presented in the same manner as other candidate characters or figures, and further, the input user selects a symbol indicating the character string. By doing so, the detailed contents are presented, so even if the device has only a small presentation surface such as a portable device, the input user can easily see the recognition candidates and easily determine the expected candidates. Will be able to do it.

When the recognition result candidate group is composed of only character strings registered in advance, the corresponding character string is immediately presented upon input of handwriting data. Can be determined more quickly and easily.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an online character / graphic recognition apparatus to which a character string presentation method according to the present invention is applied.

FIG. 2 is an explanatory diagram showing a direct correspondence between a handwriting input by a pen and a character string, and an example of registration of the character string.

FIG. 3 is a schematic diagram illustrating a presentation example of a recognition result candidate group including a symbol indicating a character string.

FIG. 4 is a schematic diagram showing an example of presenting specific contents of a character string.

FIG. 5 is a schematic diagram showing a presentation example of a plurality of character strings.

FIG. 6 is an explanatory diagram showing a direct correspondence between a handwriting input by a pen and a plurality of character strings, and an example of presenting the character strings.

FIG. 7 is a flowchart showing control for presenting a recognition result candidate by the character string presenting method according to the present invention.

FIG. 8 is a schematic diagram showing an example of presenting a character string as a recognition result candidate by a conventional character string presentation method.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Input data extraction part 2 Online recognition part 3 Recognition dictionary 4 Recognition result candidate presentation part 5 Selection part

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-6-322610 (JP, A) JP-A-5-128313 (JP, A) JP-A-4-31990 (JP, A) JP-A-1- 253079 (JP, A) JP-A-61-292774 (JP, A) JP-A-60-7513 (JP, A) JP-A-60-217487 (JP, A) (58) Fields investigated (Int. ^7, DB name) G06K 9/03 G06K 9/62 G06F 17/22 506 G06F 3/03 380

Claims (4)

(57) [Claims] 1. A character string presentation method in online recognition for recognizing an input drawing trajectory as a character, a figure, or a character string using an input device and outputting the recognition result, comprising the steps of: A determining step of determining whether or not a character string is included in the recognition result candidate group; and determining that the character string is included in the recognition result candidate group, A first presentation step of presenting in the first presentation area together with a candidate character or figure group of a recognition result; a selecting step of selecting a candidate to be determined from the presented recognition result candidate group; When the symbol indicating is selected,
A second presentation step of presenting the corresponding character string in a second presentation area. 2. In the first presenting step, when the recognition result candidate group includes two or more character strings, each of the character strings indicates a character string. The character according to claim 1, wherein a symbol is presented in a first presentation area, and in the second presentation step, a character string corresponding to the selected symbol is presented in the second presentation area. Column presentation method. 3. In the first presenting step, when two or more character strings are included in the recognition result candidate group, all of the character strings are represented as character strings. Presenting one of the symbols shown in the first presentation area, and, in the second presenting step, when the symbol indicating a character string is selected,
2. The character string presentation method according to claim 1, wherein all applicable character strings are presented in the second presentation area. 4. A method according to claim 1, further comprising a second determining step of determining whether or not the recognition result candidate group includes only a character string, wherein the recognition result candidate group includes only a character string. 2. The character string presentation method according to claim 1, further comprising a third presentation step of presenting the corresponding character string in the second presentation area when it is determined that the character string is a character string.