KR100320062B1 - Hand-written input display - Google Patents

Abstract

It is an object of the present invention to correctly recognize a character entered by hand without re-entering it.

In the configuration of the present invention, when five characters are inputted into five screens as the evaluation name, the five screens are divided into a predetermined number of character segments, and the characters are recognized for each character segmentation. Whenever the second screen is recognized as one character along with the third screen whenever the original screen is recognized to constitute the same characters as the first screen, a division indicating the division between the second screen and the third screen. The indicator 51 (gesture) 51 is input by hand. This divides each screen constituting different characters into different character segments. As a result, it becomes possible to recognize them as correct characters.

Description

Hand written input display

The present invention relates to an input display device most suitable for handwriting, for example, for use in an electronic organizer device, a portable computer and the like.

Recently, a device capable of inputting letters by hand, such as an electronic notebook device, has become widespread. As such, a device that can be input by hand is different from a word processor and the like, and a character can be input by writing a letter on a board using a pen without a keyboard. You can easily enter text. Compared with the case of using a keyboard, even a beginner can easily enter a character since there is no need to remember the position of a key.

As described above, in a device capable of inputting characters by hand, there are some types of input frames for inputting characters and those capable of inputting characters at arbitrary positions without providing the input frames.

In a device provided with an input box, a user inputs letters by hand one by one in the center of the input box. This input frame is usually divided into several areas, and the user inputs one character in each area. Because the location of input is limited, the operability becomes poor by the user, but since the area for inputting each character is separated, it is not necessary to break the input characters one by one, and it is possible to recognize each character relatively accurately. .

On the other hand, in the apparatus which does not provide an input frame, a user inputs a character by hand in arbitrary positions on an input board (tablet). In this case, a character input by hand was recognized as a character at a predetermined timing.

That is, when a character is input, the boundary between the input character and the character is determined. Each broken character is recognized as one character, and the recognition result is displayed.

As a result, for example, as shown in FIG. 16 (a), when the user inputs a kanji character of "明", when the characters "日" and "月" are separated by two characters, In this case, the input character is recognized as a kanji character of "明".

Therefore, as shown in FIG. 16 (b), when the characters of the letters "day" and "month" are relatively spaced apart, these characters are separated as separate characters. Each character of "Mon" is recognized as a separate character.

In the conventional apparatus, if the result recognized as the character of "明" or the result recognized as the characters of "day" and "month" is different from the user's intention, the user must input again and input the intended character. There was a problem.

This invention is made in view of such a situation, and is made to input a desired character, a figure, etc. more quickly.

The hand-written input display device of the present invention comprises input means for inputting handwriting by hand (for example, pen 16 of FIG. 1) and recognition means for recognizing handwriting inputted by the input means (for example, Program step S1 of FIG. 2), display means (e.g., display 13 in FIG. 1) for displaying the handwriting inputted by the input means and the recognition result obtained by the recognition means, and the handwriting. It is characterized by including a changing means (e.g., split indicator 51 in Fig. 4) for changing the writing of several handwriting in the case of recognition.

This apparatus is further provided with selecting means (e.g., the candidate rank display section 61 in FIG. 8) for selecting a predetermined one of the displayed multiple recognition results when displaying the multiple recognition results on the display unit 13. As shown in FIG. Can be.

In the hand-written input display device of the above structure, when a predetermined handwriting is input, a recognition result obtained by recognizing this is displayed. When the displayed recognition result is different from the intended one, the handwriting of the input and display may be divided or combined by inputting a predetermined gesture. Therefore, it is possible to input the intended character without re-entering a new handwriting.

(Example)

1 is a block diagram showing the configuration of an embodiment of a handwritten input display device of the present invention. The CPU 1 controls each part in accordance with a program stored in the ROM 2, and executes a character input operation. The RAM 3 stores predetermined data and the like as necessary. The backup RAM 4 is always supplied with predetermined power from the battery 5, and stores information stored and stored even when the device is powered off. The timer 6 is always operated on the battery 5 on the basis of the crystal oscillator 7.

The display controller 11 expands the data input via the bus line 8 to the V-RAM 12, and displays the image developed thereon on the display unit 13. The input tablet 15 outputs the coordinate position data input by operating the pen 16 to the tablet interface 14, and the tablet interface 14 outputs the data via the bus line 8 to the CPU 1 or the display. It outputs to the controller 11.

The microphone 18 outputs the collected phoneme signal to the individual recording processing circuit 17 so as to be stored. The audio signal output from the individual recording processing circuit 17 is supplied to the speaker 21 via the adder 20 so as to be soundproofed. The dialer 19 is controlled by the CPU 1 and outputs a predetermined dial signal to the speaker 21 via the adder 20.

Several key switch groups 22 are operated when switching the operation of the device, and a signal corresponding to the operation is output from the key interface 23 to the CPU 1 via the bus line 8. In addition, by operating a predetermined number of key switch groups 24, the power supply circuit 25 is turned on to supply power to each part, or the recording, reproducing, stopping, volume adjustment, etc., by the individual recording processing circuit 17 are performed. You can run it.

An IC card (not shown) can be inserted into the socket 26, and the IC card interface 27 is configured to exchange data with the IC card.

The pin connector 29 can be connected to, for example, a modem, and can transmit and receive a signal via the expansion interface 28 and the modem.

In addition, the input tablet 15 is made of a transparent material and formed on the display unit 13. As a result, the display of the display unit 13 can be viewed through the input tablet 15.

Next, with reference to the flowchart of FIG. 2, the operation | movement in the case of inputting by hand by character is demonstrated. When the pen 16 writes a character on the input tablet 15 by hand, the handwriting data is supplied to and stored in the handwriting handwriting database DBI of the RAM 3 via the tablet interface 14. This handwriting data is supplied to the V-RAM 12 by the display controller 11 and developed as bitmap data. The bitmap data is read out by the display controller 11 and output to the display unit 13 for display.

As a result, the handwriting input by the pen 16 is displayed on the display unit 13.

In step S1, the CPU 1 reads the handwriting data stored in the handwriting handwriting database DBI in this manner and executes the character recognition processing. The data obtained as a result of the recognition is supplied to and stored in the recognition result database DB2.

Next, the flow advances to step S2, and the first recognition result having the highest score (high similarity) among the recognition results stored in the recognition result database DB2 is read out and displayed on the display unit 13.

3 schematically shows the above operation. In other words, if a character of "い し い" is input in the evaluation name now, handwriting data of this character is subjected to character recognition processing, and in this embodiment, three candidates are obtained.

That is, in the evaluation name, the character of "い し い" is composed of five screens, but in the first candidate, the five screens are divided into four character segmentation. One character segmentation is formed by the first screen, and this is recognized as the evaluation name "し". One character segmentation is formed by the second screen and the third screen. In this embodiment, it is recognized as " L " In addition, the fourth screen and the fifth screen are recognized as different character segmentation, respectively, and the fourth screen is recognized as the number "1" and the fifth screen as ")".

In the second candidate, like the first candidate, five screens are divided into four character segmentations. The first character segmentation is composed of the first screen and the second screen, and is recognized as "い" of the evaluation name. The second character segmentation is constituted by the third screen and is recognized as "SHI" of the evaluation name. The third and fourth character segments are each composed of a fourth screen and a fifth screen, and are recognized as the numeral " 1 " and the symbol ") as in the case of the first screen.

In the third candidate, five screens are divided into three character segmentations. The first character segmentation is composed of the first screen and the second screen, and is recognized as "い" of the evaluation name. The second character segmentation is constituted by the third screen and recognized as the evaluation name "し". The third character segmentation is composed of a fourth screen and a fifth screen, and is recognized as "い" in the evaluation name.

In this embodiment, the handwriting input among the three candidates described above and the first candidate as the character recognition result are displayed on the display unit 13.

In addition, the recognition result is information related to the character segmentation (for example, the coordinates of a rectangle (circular rectangle) circumscribed to the handwriting, data on the screen constituting the character, etc.). It has a similarity (score) for the characters it produces.

It is not necessary for the circumscribed rectangle for each character segmentation to display this, but it becomes easy to input correction of the character segmentation mentioned later.

Returning to FIG. 2, the process proceeds to step S3 after step S2, and accepts a new input from the user. In step S4, it is determined whether or not the input received in step S3 instructs correction of the character segmentation. If the input handwriting does not instruct the modification of the character segmentation, the process ends. If the input received in step S3 indicates correction of character segmentation, the flow advances to step S5 to determine the recognition result corresponding to the correction instruction again.

That is, the CPU 1 accesses the recognition result database DB2 and corrects the recognition process again in accordance with the correction instruction of the character segmentation instructed by the user in step S3. The recognition result obtained by the correction is supplied to the correction recognition result database DB3 and stored.

Correction of the character segmentation is performed as shown in Figs. 4 and 5, for example.

In the embodiment of Fig. 4, the first candidate for the processing result of step S2 is displayed with handwriting as the character recognition result. The user inputs the character of "い し い" in the name of the evaluation, or the first screen and the second screen constituting the character "い" in the handwriting are not recognized as one character, and the second screen among them is an independent one character. Each time it is recognized, it is recognized to form a third screen ("し") and one character. Thus, the user instructs to divide the character segmentation composed of the second screen and the third screen as shown in FIG. An approximately V-shaped segmentation indicator (gesture) 51 is input.

When the CPU 1 receives the input of the division indicator 51, the CPU 1 divides one character segmentation constituted by the second screen and the third screen into two character segmentations. That is, the second screen and the third screen each perform a recognition process that is comparable to compose different characters.

In the embodiment of FIG. 5, since the fourth screen and the fifth screen, which should originally constitute one character, are different character segments, the combination indicator (gesture) 52 for combining the two character segments as one character segmentation is shown. The state of inputting is shown. The combining indicator 52 is a symbol that causes the approximately V-shaped symbol to fall horizontally across each of the several character segments to be combined.

When such an input is made, the CPU 1 processes the multiple character segmentation that spans this combining instruction 52 as one character segmentation.

FIG. 6 shows a detailed description of the process of obtaining the recognition result corresponding to the correction instruction of step S5 in FIG. As shown in the same drawing, it is first determined whether or not the correction instruction of the character segmentation input in step S31 indicates the segmentation combination. For this embodiment the modification of the character segmentation is either division or combination.

If the input now does not indicate a defect in the character segmentation, that is, indicative of the division of the character segmentation (in the case where the division indicator 51 shown in FIG. 4 is input), the process proceeds to step S35. The segmentation position of the segmentation to be segmented is obtained from the coordinate values of the circumscribed rectangle of the handwriting segmentation symbol (dividing gesture).

In the case of the embodiment of Fig. 4, since the character segmentation intersecting the circumscribed rectangle of the division indicator 51 is the second character segmentation, it is selected as the character segmentation to be divided. Since the division instruction symbol 51 is located between the second screen and the third screen, a division position is determined between the second screen and the third screen.

Subsequently, the procedure proceeds to step S36, where a result example with segmentation divided before and after the segmentation position of the segmentation obtained in step S35 is obtained from the recognition result database DB2. That is, in this embodiment, since the handwriting of the second screen and the third screen is processed to constitute different characters (character segmentation), the handwriting of the first screen and the second screen is regarded as one character "い". It is recognized. The third screen is also recognized as one character "し".

The process proceeds to step S37, and the result obtained in step S36 is supplied to and stored in the correction recognition result database DB3. That is, the third candidate in FIG. 3 is stored in the correction recognition result database DB3.

On the other hand, if it is determined in step S31 that the gesture inputted is the engagement indicator 52 for instructing engagement, the flow advances to step S32. The set of segmentation to be combined is obtained from the coordinate values of the eccentric square of the handwriting.

In the case of the embodiment of Fig. 5, since the combining indicator 52 is input over the character segmentation constituted by the fourth and fifth screens, the character segmentation is performed whenever these character segments are processed as one character segmentation. Is selected as.

The flow advances to step S33, and a result string including a result of the set of character segmentation obtained in step S32 in the recognition result database DB2 being one character segmentation is obtained. In the case of the embodiment of FIG. 5, the character "i" obtained by recognizing the fourth screen and the fifth screen as one character is obtained.

In addition, the flow advances to step S34, and the result obtained in step S33 is supplied to the correction recognition result database DB3 and stored. In the case of the embodiment of FIG. 5, the character "i" is stored in the correction recognition result database DB3.

In FIG. 2, the process of integrating the recognition result corresponding to the correction instruction of step S5 is performed as described above, and then proceeds from step S5 to the process of step S6 or less, and there are several recognized results. If so, a predetermined one is displayed.

In step S6, the internal variable j is initially set to one. The flow advances to step S7 to display the j-th recognition result on the display unit 13. That is, the CPU 1 reads out the first (j = 1) of the data stored in the step S5 from the correction recognition result database DB3, outputs it to the display unit 13, and displays it. The flow then advances to step S8 to accept new input from the user. In step S9, it is determined whether or not the handwriting accepted in step S8 instructs to display the next candidate. If it is not the next candidate display instruction, the flow advances to step S13 to determine whether or not the display of the previous candidate is instructed.

If it is determined in step S9 that the next candidate display instruction is made, the process proceeds to step S10, where the variable j is increased by one and the process proceeds to step S11. In step S11, it is determined whether or not the variable j increased in step S10 is larger than the number of candidates stored in the recognizing recognition result database DB3. If it is determined that the variable j is less than or equal to the number of candidates stored, the candidates to be displayed next remain, so that the process returns to step S7 and the recognition result corresponding to the variable j increased in step S10. Display. That is, the next candidate is displayed.

On the other hand, when it is determined in step S11 that the variable j is larger than the number of candidates stored in the correction recognition result database DB3, that is, when the last candidate is already displayed, the process proceeds to step S12. Then, the variable j is set to 1, the flow returns to step S7, and the first recognition result is displayed. That is, the display of the candidate of the recognition result is changed to the next candidate in turn, and when the candidate is finally reached, the candidate is returned to the first candidate.

On the other hand, when it is determined in step S13 that the handwriting input in step S8 indicates the display of the previous candidate, the flow advances to step S14, and the variable j is decreased by one. If it is determined in step S15 that this variable j is one or more, the previous candidate remains as a recognition candidate, so the process returns to step S7 and the recognition corresponding to the variable j reduced in step S14. Display the result. In other words, the previous candidate for recognition is displayed.

In the case where it is determined in step S15 that the variable j is smaller than 1, since the first recognition result is displayed, there is no previous candidate in this case. In this case, the process proceeds to step S16 where a value equal to the number of candidates for the correction database is set in the variable j. That is, the last candidate is selected, and the process returns to step S7 and this is displayed.

In this way, the candidates to be displayed are sequentially changed to the front and back, and when the first candidate is reached, the display is changed to the last candidate.

If it is determined in step S13 that the handwriting accepted in step S8 does not instruct the display of the previous candidate, the process returns to step S3, and waits for a new input again, and performs processing corresponding to the handwriting. Will be done.

As described above, when there are several recognition candidates in steps S6 to S16, one of them can be selected and displayed one by one.

In addition, the input (command) in the above step S3 or step S8 can be performed by operating the predetermined position on the input tablet 15 by the pen 16 outside the gesture. Alternatively, it is possible to input from a mouse, a keyboard, and the like.

In this way, any one of the first to third candidates shown in FIG. 3 is sequentially displayed.

The above process is summarized as shown in FIG. In other words, as shown in the same drawing (a) for the first time, the handwriting of the character of "い し い" by hand is displayed. When the input handwriting is recognized, as shown in the same drawing (b), the first candidate of the recognition result is displayed at the lower right of the handwriting. The user can see the recognition result and know that the second screen and the third screen are included in one character segment whenever it is recognized as another character.

Thus, the user inputs a division indicator 51 for dividing this second character segmentation into two character segments. As a result, as shown in the same drawing (c), one character "i" is recognized and displayed on the first screen and the second screen, and one character "し" is recognized and displayed on the third screen.

In addition, since the fourth screen and the fifth screen are recognized as different character segmentations each time they are recognized as one character, inputting the combination indicator 52 that combines them causes the first screen as shown in the same drawing (d). The character "い" obtained as a result of recognizing the fourth screen and the fifth screen as one character is displayed.

Enter the number 1 by hand as described above, and then input the instruction to split or combine the character segmentation by referring to the recognition result as necessary, so that the original character can be recognized without rewriting by hand. Can be.

8 shows another display example. In this embodiment, as shown in the same drawing (a), the handwriting is displayed by inputting the character of "iro". By this handwriting, character segmentation is interrupted, and character recognition is performed for each character segmentation. As shown in the same drawing (b), an external rectangle representing character segmentation is displayed. In this embodiment, each of the first to sixth screens is divided as different character segmentation A to F. In FIG.

FIG. 8C shows a state in which the results of recognizing each such character segmentation A to F are displayed (four pairs in this embodiment). When selecting a predetermined pair from among the displayed multiple pairs, the corresponding one of the candidate rank display units 61 displayed on the left side is touched with the pen 16. As a result, the touched pair is selected as the recognized result.

In this embodiment, the scroll bar 62 is displayed on the far right side, and is displayed by touching the pen display 16 with the triangular display portion 62a in the uppermost direction or the triangular display portion 62b in the lowermost direction. When there are five or more pairs of candidates, the display state can be scrolled up and down. Touching the display 62a scrolls the candidate upwards, and touching the display 62b scrolls the candidate downwards.

9 also shows another display example. In this embodiment, as shown in the same drawing (a), the handwriting is displayed when the first to sixth screens are input. Then, the character segmentation corresponding to the recognition candidate of the first position is displayed as shown in the same drawing (b). In addition, as shown in the same drawing (c), several candidates for several character segmentation are displayed simultaneously (four pairs in this embodiment).

Even in this case, the predetermined recognition result can be selected by touching a predetermined portion of the candidate rank display section 61 with the pen 16, and the display of the candidate can be scrolled by operating the display section 62a or the display section 62b. It is the same as the case in FIG.

10 also shows another display example. In this embodiment, in the state where multiple recognition results are displayed as shown in the same drawing (a), as shown in the same drawing (b), the forward and backward display instruction gestures 71 are moved by the pen 16. The writing can be scrolled up or down by inputting or writing the next candidate display instruction gesture 72 with the pen 16, as shown in the same drawing (c). In this embodiment, these front and rear display indication gestures 71 and next candidate display instruction gestures 72 are triangular figures pointing upward or downward.

In this embodiment, instead of inputting the forward / backward display instruction gesture 71 or the next candidate display instruction gesture 72, the display can be scrolled in the same manner by touching the display portions 62a and 62b of the scroll bar 62 with a pen. It is possible.

Figure 11 shows an example of the format of a handwritten database (DBI). As shown in the same figure, the screen number is arranged first, and the number n of coordinate points of the screen corresponding to the screen number is arranged next. Then, further, coordinate values (x ₀ , y ₀ ), (x ₁ , y ₁ ) ... (x _n-1 , y _n-1 ) of n coordinate points are arranged. In the case where several screens are input in the same manner, the number, coordinate score, and coordinate value of the screen are sequentially arranged. Finally, the file terminator is placed.

12 shows an example of the format of the recognition result database DB2 and the correction recognition result database DB3. First, recognition results corresponding to each character segmentation are arranged in sequence. In the recognition result for each character segmentation, the ID of the character segmentation is arranged at the head, and the start screen number and the end screen number of the character segmentation are next placed. Next, the coordinates (x ₀ , y ₀ ) and (x ₁ , y ₁ ) of the diagonal vertex of the circumscribed rectangle of the character segmentation are disposed. Next, the number N of recognition candidates is arranged, followed by the character codes and grade values of the first to Nth recognition candidates.

At the end of the recognition result for the one-character segmentation, the end result of the recognition result for the one-character segmentation is placed. Then, the character segmentation constituting the character recognition result obtained by combining the character segmentation so that the grade value becomes high is arranged one by one. The character recognition result is composed of the candidate ID, the number M of character segments constituting the candidate, and the IDs ID0 to IDM of the character segments. And finally, a file exit code is added.

In the above description, the case of inputting a character by hand is described as an example, but the present invention can be applied to a case of inputting a figure.

13 and 14 show an embodiment in this case. In FIG. 13 (a), for example, in the case of drawing an electric block diagram, three rectangles and line segments connecting the rectangles are input and the handwriting is displayed. In the same drawing (b), the handwriting recognition result shown in the same drawing (a) is also displayed, and the figure segmentation of each figure is also displayed.

In this embodiment, the input figure is divided into five figure segmentations, but the second figure segmentation is divided including the line segments that should originally constitute the first figure. Thus, as shown in the same drawing (b), when the division mark 51 is inputted to a predetermined position, as shown in the same drawing (c), the line segments that have been included in the second figure segmentation up to that time are the first. It is changed to be included in the first figure segmentation.

Since the graphic segmentation is divided into the rectangle and the line segment connecting the same by the above process, the process may be terminated here, but several of the figure segments can also be integrated into one figure segment.

In this case, as shown in Fig. 14 (a), the combining instruction symbol 52 is input so as to cover one figure segmentation (third to fifth figure segmentation in this embodiment). Thereby, as shown in FIG. 14 (b), these 3rd-5th figure segmentation is integrated as one segmentation.

Figures integrated in one figure segmentation are processed as a whole when, for example, moving, erasing, or the like is performed.

In the case of inputting a figure as described above, in the case of multiple recognition candidates, the desired candidates can be displayed one by one. Fig. 15 shows this embodiment.

That is, in Fig. 15 (a), one input handwriting and one recognition result corresponding thereto are displayed. This recognition result can be displayed simultaneously, for example, as shown in FIG.

Then, as shown in FIG. 15 (b) or (c), the recognition result displayed by inputting the forward / backward display instruction gesture 71 or the next candidate display instruction gesture 72 is scrolled forward or backward. You can do that.

As described above, the present invention can be applied to a case in which arbitrary shapes such as letters and figures are displayed.

As described above, the handwriting input display device of the present invention is adapted to change the number of handwritings by the changing means. If the handwritten input shape is not recognized as the intended shape, it is not inputted again. It can be quickly recognized by the intended shape.

1 is a block diagram showing the configuration of an embodiment of a handwritten input display device of the present invention.

2 is a flowchart for explaining the operation of the embodiment of FIG.

FIG. 3 is a diagram illustrating the processing of step S1 and step S2 in FIG. 2.

4 is a diagram for explaining the division of character segmentation.

5 is a diagram illustrating the combination of character segmentation.

6 is a flowchart for explaining the detailed processing of step S5 of FIG.

FIG. 7 is a diagram for explaining the processing of steps S1 to S5 in FIG.

8 is a view for explaining an example of displaying a plurality of candidates for the same segmentation.

9 is a diagram for explaining display examples of several candidates for several character segments.

10 is a diagram for explaining an example in which the display scrolls.

11 is a diagram for explaining the format of a handwritten database.

12 is a diagram for explaining the format of the recognition result database.

13 is a diagram for explaining an example of changing the shape segmentation.

FIG. 14 is a view following FIG. 13. FIG.

15 is a diagram for explaining an example of scrolling the display of figures.

16 is a diagram for explaining a display example of a conventional handwritten input display device.

* Explanation of symbols for main parts of the drawings

1. CPU 2. ROM

3. RAM 8. Busline

11.Display Controller 12.V-RAM

13. Display 14. Tablet Interface

15. Input Tablet 16. Pen

51. Split indicator 52. Combined indicator

61. Candidate ranking display 62. Scroll bar

71. Front and rear display indication gesture 72. Next candidate display direction gesture

Claims (1)

Handwriting input means for inputting handwritten handwriting; Recognition means for recognizing the handwriting inputted by the input means; Display means for displaying a handwriting inputted by the input means and a recognition result obtained by dividing the handwriting into a plurality of segments as a recognition result obtained by the recognition means; If a segmentation indicator for dividing one segmentation of the plurality of segmentations displayed on the display means into two or more is inputted by a user's gesture, the segmentation corresponding to the input of the segmentation indicator is divided and the plurality of segmentation marks are divided. If a combination indicator for combining two or more neighboring segments of one of the segmentation of the input by the user's gesture, by combining the segmentation corresponding to the input of the combination indicator into one segmentation, the plurality of segments A hand-written input display device comprising a recognition result changing means for changing to segmentation of a new arrangement desired by a user.