JPH0496886A - Handwritten input recording and recognizing device - Google Patents

Abstract

PURPOSE:To improve a working efficiency by preserving characters and graphics inputted by a handwriting by feature amounts obtained at the time of inputting them, displaying the characters and the graphics, and then recognizing only character and graphic parts selected by a user. CONSTITUTION:The characters or the graphics are written in an inputting means 1 by the handwriting. A feature amount extracting part 2 operates pre- processings such as smoothing and sampling, extracts the feature amounts of the characters or the graphics, and outputs them to a feature amount storing part 3. The storing part 3 holds the feature amounts, and outputs the coordinates of feature points to a reconstructing part 4. The reconstructing part 4 reproduces approximately the characters or the graphics inputted by the handwriting, and outputs them to a display part 7. The user selects the characters or the graphics to be converted into types or fair copy graphics from among them by using a reading means 5, reads out the feature amounts of the characters and the graphics from the feature amount storing part 3, outputs them to a recognizing part 6, obtains the display coordinates of the type characters or the fair copy graphics being the recognized result, and outputs them to a display part 7.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a handwritten input record recognition device that records, recognizes, and displays handwritten characters and figures.

BACKGROUND ART FIG. 4 shows a configuration diagram of this conventional handwritten input record recognition device, where l is an input means such as a tablet, 2 is a feature extracting unit for extracting features of the input data, and 6
Reference numeral 41 indicates a recognition unit that recognizes characters and figures from the extracted features, 41 a recognition result storage unit that stores the recognition results, and 7 a display unit that displays the recognition results.

In the conventional handwritten input record recognition device configured as described above, first, the user inputs characters, figures, etc. by handwriting into the input means 1.

Next, the feature extraction unit 2 extracts feature points, stroke order, and strokes from the pen point coordinates and pen pressure when characters and figures are input.
The information such as the above is extracted as the characteristic amount of the characters and figures and outputted to the recognition unit 6.

The recognition unit 6 recognizes characters and figures handwritten by the user from the feature quantities inputted from the feature quantity extraction unit 2, encodes them, and sends the printed characters and neatly written figures as recognition results to the recognition result storage unit 41.

The recognition result storage section 41 holds the recognition results sent from the recognition section 6 and outputs the recognition results to the display section 7.

The display unit 7 displays the recognition results sent from the recognition result storage 41 as printed characters or neatly written figures.

Problems to be Solved by the Invention However, in the configuration of the conventional handwritten input record recognition device as described above, all handwritten characters and figures are always recognized, stored, and output, so even if they are not important, Even if it is only a partial part, it must be recognized, and if it cannot be recognized, it must be corrected each time, resulting in poor work efficiency. Furthermore, if there is an error in the recognition of an important part, the error may not be noticed when looking at the recognition result, which may cause trouble later on.

In view of the problems of the conventional handwritten input record recognition device, the present invention has been developed to recognize characters and figures input by hand in the conventional manner.
It does not always recognize everything and save the results, but it saves the features obtained during input, and uses those features to display and use characters and shapes without recognition. To provide a handwritten input record recognition device with high work efficiency by recognizing only characters and graphic parts selected by a person.

Means for Solving the Problems The present invention provides an input means for inputting characters and figures by handwriting by a user, and extracts feature quantities of the characters and figures from information on the characters and figures input from the input means. A feature extraction unit, a feature storage unit that stores the features extracted by the feature type extraction unit, and a feature storage unit that composes handwritten characters and figures from the features stored in the feature storage unit. a reconstruction unit for displaying on the first display unit; a reading unit for reading out feature quantities of characters and figures selected by the user from the feature quantity storage unit; and a reading unit ζ. This is a handwritten input record recognition device that recognizes characters and figures input from feature amounts and outputs the recognition results on a second display unit.

Effect of the present invention With the above-described configuration, the present invention calculates feature quantities such as feature points, stroke order, and strokes from information such as pen point coordinates and pen pressure obtained when a user inputs characters or figures by hand into an input board. After the user completes all input work, the user can perform the work more efficiently by recognizing only the characters and figures that he or she wants to process, such as fair copying.

Embodiments Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 shows a configuration diagram of a handwritten input record recognition device according to an embodiment of the present invention. In FIG. 1, 1 is an input means for the user to input characters and figures by hand, and 2 is a feature for extracting the feature quantities of the characters and figures from the information on the characters and figures input from the input means 1. 3 is a feature storage unit that holds the feature φ extracted by the width extraction unit 2; 4 is the feature ff;
From the feature amounts held in the storage unit 3, characters and figures input by hand are constructed in a nearly 11-dimensional manner, and displayed on the first display unit 7.
5 is a reading means for reading out the feature quantities of characters and figures selected by the user from the feature Φ storage section 3; 6 is a reading means from the feature quantities read out by the reading means 5; This is a recognition unit that recognizes the input characters and figures and outputs the recognition results to the second display unit 7 (in this embodiment, it also serves as the first display unit 7).

The operation of the handwritten input record recognition device according to the embodiment of the present invention configured as described above will be described below.

First, the user writes characters and figures by hand onto the input means l. The input means 1 may be, for example, a data tufflet. A data tablet consists of a tablet that serves as a writing surface and a stylus that serves as a pen.
The positional relationship between the tablet and the stylus is detected by using, for example, a sheet made of piezoelectric material stretched over the surface of the tablet, and an electrical signal generated by the force exerted by the stylus. By outputting data using a unique clock, the data tablet can extract the coordinates of the writing point as a time series, and can also extract information on whether or not the tip of the writing brush is touching the writing surface.

Next, the feature extracting section 2 first performs preprocessing such as smoothing and sampling in order to absorb fluctuations in the input sent from the data tablet 1, for example.

As a smoothing process, in order to remove noise with a large fluctuation range, if temporally adjacent coordinates are separated by a certain distance H(It) or more, that coordinate is replaced with the previous coordinate, and An averaging operation is performed on temporally adjacent coordinates to remove noise with small fluctuations. As a sampling process, in order to eliminate the effects of writing speed and other redundant coordinate sequences, we use a fixed distance between temporally adjacent coordinates (12).
If the distance is less than or equal to the distance, the coordinates are removed.

The sampling process equalizes the speed of the pen point movement, and since the effects of writing speed and redundant coordinate series are removed, the length of the obtained pen point coordinate series is equal to the length of the input character. It corresponds to the total length of the lines of the figure.

Next, a series of writing point coordinates from the time when the tip of the brush touches the writing surface until the time when it leaves the writing surface is extracted as a stroke that constitutes a character or figure, and the writing points that divide the stroke into N equal parts are taken as feature points. For example, by matching the pen point coordinate series with a standard series possessed by the feature extracting unit 3 from the beginning, it is classified as a basic stroke constituting the character or figure. Furthermore, (1) the coordinates of the feature points, (2) the type and number of the basic strokes, and (3) the order of appearance (order of strokes).
, (4) Mutual positional relationship, (5) The mutual positional relationship and appearance order of partial patterns composed of several basic strokes, or the partial patterns and basic strokes, are extracted as feature quantities of the characters and figures. , is output to the feature quantity storage section 3.

The feature amount storage section 3 holds the feature amounts sent from the feature amount extraction section 2, and also stores the feature amounts sent from the feature amount extraction section 2 to the reconstruction section 4, for example.
) outputs the coordinates of the feature points shown in ).

The reconstruction unit 4 reconstructs the pen point coordinate series of characters and figures input by hand by the user by performing interpolation using a spline function or linear interpolation between the coordinates of the feature points sent from the feature storage unit 3. Then, the characters and figures are approximately reproduced and output to the display section 7.

The user uses the reading means 5 to select characters and figures that he wishes to convert into printed characters and neat figures from among the characters and figures reconstructed by the reconstruction unit 4 and displayed on the display unit 7. That is, the reading means 5 is, for example, a transparent data tablet 21 pasted on the display section 7 and integrated with the display section 7, as shown in FIG. As shown in FIG. 2(a), the user first presses the mote switching key 22 on the tablet 21 to set it to conversion mode. Next, as shown in FIG. 6(b), use the stylus 24 to surround the character or figure you want to convert with a closed curve, and then press the conversion key 23. The amount of data read from the quantity storage section 3 is output to the amount recognition section 6. If the data tablet 31 and the display section 7 are independent as shown in FIG. 3, the user can press the mote switching key 32 on the tablet 31 and switch the By doing so, the display unit 7 displays a display 3 indicating the current pen point coordinates of the stylus 24.
3 is issued, the character or figure to be converted is surrounded by a closed curve using the display 33 indicating the writing point coordinates of the stylus 24, and the conversion key 34 is pressed as shown in FIG. Feature quantities of characters and figures are stored in the feature quantity storage unit 3.
, and outputs it to the recognition unit 6.

The recognition unit 6 measures the Euclidean distance from the feature quantity sent from the reading means 5, for example, to the standard time minute amount of the printed characters and clean-cut figures that the recognition unit 6 has from the beginning, and determines the shortest printed characters and clean-cut figures. By using the figure as a recognition result, the user can recognize whether the character or figure that he or she wants to convert is a typed character or a fine-printed figure, and from the position coordinates of the character or figure that the user wants to convert, the user can recognize the typed character or figure that is the recognition result. The display coordinates of the printed figure are determined and output to the display section 7.

The display unit 7 refers to the coordinate points sent from the reconstruction unit 4 and the recognition unit 6, and displays the coordinate points on the display unit 7 as shown in FIGS. 2(c) and 3(c). do.

As described above, according to the present embodiment, by providing the feature amount storage section 3, the reconstruction section 4, and the reading means 5, the pen point obtained when the user inputs characters or figures by hand on the input board Characteristics such as feature points, stroke order, and strokes are recorded from information such as coordinates and pen pressure, and after the user has finished all input work, the characters input by hand by the reconstruction unit 4 are recorded. The user can work more efficiently by rearranging the characters and figures and recognizing only the characters and figures that the user wants to process, such as fair copying. In other words, you can see roughly reconstructed approximate characters and figures without going through the wasteful and troublesome process of recognizing all unnecessary parts, making it easier to detect input errors. (There is no need to worry about seeing incorrect recognition results like in the past).

Note that although a data tablet and a stylus are shown as input means as examples, a mouse or the like can also be used to obtain the pen point coordinates, so that similar actions and effects can be obtained.

As described in detail, the present invention does not recognize all handwritten characters and figures, but only those selected by the user, making the recognition process more efficient and reducing work. It is possible to provide an efficient handwritten input record recognition device.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a handwritten input record recognition device according to an embodiment of the present invention, and FIG. 2 is an operation diagram when the display section and data tab of the embodiment are integrated.
FIG. 3 is an operational diagram when the display unit and data tablet of the same embodiment are independent, and FIG. 4 is a block diagram showing the configuration of a conventional handwritten input recording recognition device. DESCRIPTION OF SYMBOLS 1... Input means, 2... Display part, 3... Feature quantity extraction part, 4... Feature quantity marking part, 5... Reconstruction part, 6.
... Reading means, 7... Recognition unit, 21.31...
Data tablet, 24...Stylus, 22.32
...Mode switching key, 23.34...Conversion key 33...Display showing the current pen point coordinates of the stylus,
41...Recognition result storage unit. Agent Patent Attorney Matsu 1) Tadashi Michi Figure 2 (c) Figure 3

Claims (2)

[Claims] (1) an input means for a user to input characters and figures by hand; a feature quantity extraction unit for extracting the characteristic quantities of the characters and figures from the text and figure information input from the input means; A feature amount storage section that holds the feature amount extracted by the feature amount extraction section, and a handwritten input character or figure from the feature amount held in the feature amount storage section, and a first display. a reconstruction unit for displaying a character or figure selected by the user in the feature storage unit; A handwritten input record recognition device comprising a recognition section that recognizes characters and figures and outputs recognition results to a second display section. (2) The handwritten input record recognition device according to claim 1, wherein a single display unit is used as the first display unit and the second display unit.