JP2004139281A - Online handwritten information recognition device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an online handwritten information recognition device and method capable of easily selecting only targeted handwriting even when handwriting to be selected and any handwriting other than the handwriting to be selected closely exist. <P>SOLUTION: In this online handwritten information recognition method, handwritten coordinate data are inputted, and character recognition or graphic recognition is carried out to the group of stroke data formed of the inputted handwritten coordinate data, and the sub-set of the stroke data being the target of handwriting recognition is selected from the group of the stroke data, and the character recognition or graphic recognition is carried out to the selected sub-set of the stroke data, and the recognition result is outputted. Then, a recognition target category being the target of recognition is set, and whether or not the respective stroke data belong to the currently set recognition target category is decided, and the stroke data to be selected are limited only to the stroke data whose belonging to the recognition target category is decided. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an online handwritten information recognition apparatus and method for performing character recognition or figure recognition while recognizing information on handwritten strokes.
[0002]
[Prior art]
With the recent rapid advances in computer technology, the processing speed of CPUs has also been improved, and various types of handwritten character recognition technologies that have conventionally been considered difficult to put into practical use due to problems such as the processing time and recognition accuracy have been considered. At present, portable information terminals are rapidly spreading not only for business use but also for general home use. Recently, many technologies have been developed that can recognize handwritten characters or handwritten figures freely written without being restricted by a character frame or the like.
[0003]
When writing freely without being restricted by a character frame or a ruled line, various contents such as a figure and a character string are freely written. In order to determine the category of a character string or a figure by freely recognizing a part of the handwritten content, an operation of selecting a part to be recognized from many pieces of stroke data is required.
[0004]
For example, as shown in FIG. 1, a character string "Tanaka" is selected from the contents of a memo in which a character string and a line drawing are mixed, and character recognition is performed by acquiring stroke data thereof. Obtaining the result character string is the basic flow of the handwritten character recognition processing. In order to execute such processing, processing such as selection of stroke data (handwriting), acquisition of stroke data (handwriting), and recognition of a character string are required. Here, the stroke data means trajectory data from when the pen touches down to when it comes off.
[0005]
FIG. 2 shows a typical configuration of a conventional online handwritten information recognition apparatus. In FIG. 2, stroke (handwriting) coordinate information obtained from the coordinate input unit 21 is first stored in the stroke data storage unit 22 as a set of stroke data. Then, a part of the stroke data is selected by the stroke data selection unit 23 from the stored set of stroke data, and a subset of the stroke data is obtained. Finally, by sending a subset of the stroke data to the handwritten information recognition unit 24, a desired recognition result can be obtained.
[0006]
Specifically, the coordinate input unit 21 uses an input medium, such as a tablet or a stylus pen, capable of drawing and writing characters. The stroke data storage unit 22 may be a storage device of a computer or a storage device on a network.
[0007]
Furthermore, as the stroke data selection unit 23, it is conceivable to use a method of selecting by an encircling gesture by writing with a pen, for example, as disclosed in (Patent Document 1) or (Non-Patent Document 1). The handwritten information recognizing unit 24 also includes a normal online handwritten character recognizing means as disclosed in, for example, (Patent Document 2), (Non-Patent Document 2), and (Non-Patent Document 5), It is conceivable to use a figure recognizing unit as disclosed in US Pat.
[0008]
[Patent Document 1]
JP-A-6-131111
[Patent Document 2]
JP-A-10-40337
[0009]
[Non-patent document 1]
Satoshi Sato and two others, "Implementation and Evaluation of Object Judgment Algorithm by Pen Enclosure for Handwriting Interface", IEICE Technical Report, December 1991 (PRU92-88)
[0010]
[Non-patent document 2]
Hiroshi Tanaka, et al., "Online Handwritten Character String Recognition without Real-Time Frame Using Hierarchical Delay Segmentation", IEICE Technical Report, March 2002 (PRMU 2001-264)
[0011]
[Non-Patent Document 3]
Masashima Hiroshi, 3 others, "Online Handwritten Figure Recognition Method Independent of Number of Strokes, Stroke Order, Rotation, and Separation", Transactions of Information Processing Society of Japan, May 1986, Vol. 27, No. 5,
[0012]
[Non-patent document 4]
Haruhiko Kojima, Toru Toida, "Online Handwritten Figure Recognition by Neighboring Line Structure Analysis", Transactions of Information Processing Society of Japan, August 1987, Vol. 28, No. 8,
[0013]
[Non-Patent Document 5]
Masayoshi Okamoto, 3 others, "Handwritten Information Recognition Based on Knowledge Base", Information Processing Society of Japan 43rd (late 1991) National Convention, 2D-1, p. 2-19-2-20
[0014]
[Problems to be solved by the invention]
However, in the case where the actually written state is, for example, as shown in FIG. 3, there is a double underline which is a handwriting that is not selected, near the character string “Tanaka” to be selected. However, there is a possibility that a handwriting that is not selected is erroneously selected. In this case, it is difficult to accurately perform character recognition by also using a double underline that is not selected as a recognition target. There was a problem.
[0015]
Similarly, when the actually written state is, for example, as shown in FIG. 4, when another character string “mountain” not to be selected exists near the character string “Tanaka” to be selected. However, there is a possibility that even the stroke data of a part of the character string “mountain” that is not selected is erroneously selected, and also in this case, it is difficult to accurately perform the character recognition.
[0016]
The present invention has been made to solve the above-described problem, and even when handwriting to be selected and handwriting not to be selected exist close to each other, an on-line system capable of easily selecting only a desired handwriting can be selected. An object of the present invention is to provide a handwritten information recognition apparatus and method.
[0017]
[Means for Solving the Problems]
In order to achieve the above object, an online handwritten information recognition apparatus according to the present invention includes a coordinate input unit for inputting handwriting coordinate data, and a set of stroke data formed by the input handwriting coordinate data. A handwriting information recognition unit that performs recognition or graphic recognition, and a stroke data selection unit that selects a subset of stroke data to be subjected to handwriting recognition from a set of stroke data. An online handwritten information recognition device that performs character recognition or figure recognition in a handwritten information recognition unit and outputs a recognition result, and a recognition target setting unit that sets a recognition target category to be recognized in the handwritten information recognition unit; A strobe for determining whether each stroke data belongs to the currently set recognition target category. And Kudeta determination unit, the stroke data to be selected in the stroke data selecting unit, and having a stroke data selection targeting unit to limit the stroke data which are determined to belong to the recognition object category.
[0018]
With this configuration, by setting the recognition target category before the start of the stroke data selection process, stroke data having a different recognition target category can be excluded from the recognition target, and the handwriting to be selected and the handwriting to be selected can be excluded. It is possible to easily select only the target handwriting even when the handwriting exists nearby.
[0019]
Also, the online handwritten information recognition device according to the present invention has a group formation unit that classifies a plurality of stroke data into groups, and the stroke data determination unit determines a recognition target category to which the stroke data belongs in group units. Is preferred. Further, it is preferable that the online handwritten information recognition apparatus according to the present invention has a group formation unit for classifying a plurality of stroke data into groups, and the stroke data selection unit preferably selects stroke data for each group.
[0020]
By setting the recognition target category before the start of the stroke data selection process and by grouping the stroke data in advance, stroke data included in a group having a different recognition target category can be excluded from the recognition target. Even if handwriting to be selected and handwriting not to be selected are close to each other, and the recognition target categories of both match, it is possible to easily select only the desired handwriting. is there.
[0021]
Further, in the online handwritten information recognition apparatus according to the present invention, it is preferable that the grouping unit classifies the stroke data into groups based on the order in which the stroke data is written. Alternatively, it is preferable that the grouping unit classifies the stroke data into groups based on time information including at least a time interval at which the stroke data is written. This is because it is possible to determine whether or not a series of characters or graphics according to the length of the off-stroke.
[0022]
Further, in the online handwritten information recognition apparatus according to the present invention, it is preferable that the grouping unit classifies the stroke data into groups based on a result of character recognition of the stroke data. This is because the stroke data can be grouped for each category classified based on the evaluation value output as the recognition result.
[0023]
Further, the present invention is characterized by software that executes the functions of the online handwritten information recognition device as described above as processing steps of a computer, and specifically, a process for inputting handwriting coordinate data, A step of performing character recognition or figure recognition on a set of stroke data formed by the input handwriting coordinate data; and a step of selecting a subset of stroke data to be subjected to handwriting recognition from the set of stroke data. An online handwritten information recognition method for performing character recognition or graphic recognition on a subset of selected stroke data and outputting a recognition result, comprising: setting a recognition target category to be recognized; A step of determining whether or not each stroke data belongs to the set recognition target category; The stroke data, characterized in that it is a line handwritten information recognition method, and computer-executable program for implementing such a process and a step to limit the stroke data which are determined to belong to the recognition object category.
[0024]
With this configuration, by loading the program on a computer and executing the program, the recognition target category is set before the start of the stroke data selection process, thereby excluding stroke data having a different recognition target category from the recognition target. Even if handwriting to be selected and handwriting not to be selected exist close to each other, an online handwriting information recognition device with high recognition accuracy capable of easily selecting only the desired handwriting is provided. It can be realized.
[0025]
BEST MODE FOR CARRYING OUT THE INVENTION
(Embodiment 1)
Hereinafter, an online handwritten information recognition apparatus according to a first embodiment of the present invention will be described with reference to the drawings. FIG. 5 is a configuration diagram of the online handwritten information recognition device according to the first embodiment of the present invention.
[0026]
5, reference numeral 51 denotes a coordinate input unit, 52 denotes a stroke data storage unit, 56 denotes a stroke data selection unit, and 57 denotes a handwritten information recognition unit. It corresponds to the section 21, the stroke data storage section 22, the stroke data selection section 23, and the handwritten information recognition section 24, respectively.
[0027]
Usually, a handwriting information acquisition device combining a tablet and a stylus pen is often used for the coordinate input unit 51. Generally, the tablet senses the position coordinates of the pen at a cycle of about 50 to 200 times per second, and acquires coordinate point data including information on whether the stylus pen is being written (on state) or not (off state). I do. Then, the acquired coordinate information is stored in the stroke data storage unit 52 as a set of a plurality of stroke data as a set of one stroke data, with a continuous coordinate point sequence in which the stylus pen is on as one stroke data. In addition, as a data format of the stroke data set stored in the stroke data storage unit 52, for example, a format as shown in FIG. Of course, it is not limited to such a data format.
[0028]
Next, reference numeral 53 denotes a recognition target setting unit for setting a recognition target category in the handwritten information recognition unit 57. The recognition target categories set here include not only coarse categorization such as character strings and figures, but also fine categorization such as numeric strings, kanji strings, and polygons.
[0029]
That is, the stroke data selection unit 56 selects only the handwriting to be recognized by the handwritten information recognition unit 57. For example, if a character string is to be recognized, there is no need to select a handwriting corresponding to a figure, and if a number is to be recognized, there is no need to select a handwriting corresponding to a kanji.
[0030]
For this purpose, for example, as shown in FIG. 7, when a character string is to be selected and recognized, the recognition target category is set to “character string” in the recognition target setting unit 53, and is represented by a solid line. Only the character string that is a selectable handwriting is selected by the stroke data selection unit 56. Further, double underlines and abbreviations “Tel”, which are unselectable handwritings represented by broken lines, are excluded from selection targets in the stroke data selection unit 56.
[0031]
The recognition target category set in the recognition target setting unit 53 is passed to the stroke data determination unit 54, and from the set of stroke data stored in the stroke data storage unit 52, which stroke data matches the recognition target category It is used to determine whether or not to do.
[0032]
That is, in the stroke data determination unit 54, from among the set of stroke data stored in the stroke data storage unit 52, only the stroke data that matches the recognition target category can be selected. Determine whether the category matches.
[0033]
In the first embodiment, the stroke data determination unit 54 determines whether each of the stroke data is the stroke data forming the character string or the stroke data forming the graphic, and to which category the stroke data belongs. It shall be determined. For example, in (Non-Patent Document 5), a method of classifying a set of stroke data into four types of characters, figures, tables, and edit symbols is disclosed. By regarding the three symbols collectively as a graphic, it is possible to determine a character string and a graphic. In addition, the determination method of both is not limited to this.
[0034]
Reference numeral 55 denotes a stroke data selection target limiting unit. The stroke data determination unit 54 determines, from the stroke data determined to belong to the recognition target category, that is, from among the stroke data determined to be selected as the recognition target, Is selected.
[0035]
For example, the stroke data selection unit 56 is controlled to select stroke data surrounded by a gesture 81 (dotted arrow) written with a stylus pen or the like as shown in FIG. That is, whether or not each piece of stroke data is to be recognized is determined based on whether or not each stroke data is within an area surrounded by the gesture 81 described on the screen.
[0036]
Specifically, as shown in FIG. 9, stroke data (thick line) determined to match the recognition target by the stroke data determination unit 54 and stroke data (thin line) determined not to match are recognized in advance. In the stroke data selection unit 56, when an encircling gesture is written, even if it is stroke data included in the encircling gesture, the stroke data indicated by a thin line, that is, the stroke data Stroke data determined to not match is excluded from selection targets. Then, only stroke data indicated by a thick line, that is, stroke data determined to match the recognition target by the stroke data determination unit 54 is selected.
[0037]
Note that the stroke data selection target limiting unit 55 only needs to limit the selection target of the stroke data, so that it is not necessary to display stroke data that can be selected to the user. However, by indicating to the user which stroke data on the screen is currently selected, the usability of the stroke data selection unit 56 is considered to be further improved.
[0038]
As a method for indicating to the user which stroke data is to be selected, a method of displaying the display color of the stroke data other than the selection lighter, a method of changing the line thickness, or a method of selecting Sometimes it is possible to temporarily erase the screen. Of course, it is not limited to these methods.
[0039]
Then, the stroke data selected by the stroke data selection unit 56 is passed to the handwritten information recognition unit 57 as a subset of the stroke data, and character recognition or figure recognition is performed.
[0040]
Next, the flow of processing of a program for realizing the online handwritten information recognition apparatus according to the first embodiment of the present invention will be described. FIG. 10 is a flowchart of a program for realizing the online handwritten information recognition apparatus according to the embodiment of the present invention.
[0041]
In FIG. 10, first, the position coordinates of the pen are acquired, and coordinate point data is acquired including information on whether or not the stylus pen is being written (on state) (off state) (step S1001). Then, based on the acquired coordinate information, stroke data is extracted and stored as a set of stroke data (step S1002).
[0042]
Next, a recognition target category is set (step S1003), and it is determined whether or not individual stroke data in the stored set of stroke data matches the recognition target category (step S1004).
[0043]
If the stroke data matches the recognition target category (step S1004: Yes), it is determined that the stroke data is a selection target (step S1005), and if the stroke data does not match the recognition target category (step S1004). : No), it is determined that the stroke data is not a selection target (step S1006). Then, the processing from step S1004 is repeated until the determination is completed for all the stroke data (step S1007: No).
[0044]
If the determination has been completed for all the stroke data (step S1007: Yes), stroke data to be recognized is selected from among the stroke data that has been determined to be selected (step S1008). Finally, character recognition or figure recognition is performed on the selected stroke data (step S1009).
[0045]
As described above, according to the first embodiment, by setting the recognition target category before the start of the stroke data selection process, stroke data having a different recognition target category can be excluded from the recognition target. Even when handwriting to be selected and handwriting not to be selected exist close to each other, it is possible to easily select only the desired handwriting.
[0046]
(Embodiment 2)
Hereinafter, an online handwritten information recognition apparatus according to a second embodiment of the present invention will be described with reference to the drawings. FIG. 11 is a configuration diagram of the online handwritten information recognition device according to the second embodiment of the present invention. As shown in FIG. 11, since the basic configuration is the same as that of the first embodiment, the portions having the common functions are denoted by the same reference numerals as those in FIG. 5, and the detailed description is omitted.
[0047]
The second embodiment is different from the first embodiment in that it has a group formation unit 111 for grouping a set of stroke data stored in the stroke data storage unit 52.
[0048]
In other words, by recognizing categories to be recognized regarding stroke data such as character strings and figures, it is possible to improve recognition accuracy for handwritten information in which both characters and figures are described in the vicinity. As shown in the example, when handwritten information belonging to the same recognition target category such as a plurality of character strings or figures is described in the vicinity, it is sufficient to solve the problem that some of them are mixed and selected. Can not be solved.
[0049]
Therefore, in the second embodiment, stroke data belonging to the same recognition target category are grouped as one series as a series of character strings or a series of graphics written continuously, and the recognition target category of the stroke data is determined. By performing the stroke data selection processing on a group basis, it is possible to prevent stroke data belonging to different groups from being simultaneously selected.
[0050]
For example, as shown in FIG. 12, when character strings such as “Tel to Mr. Tanaka”, “Sama Yamamoto”, and “012-345-6789” are grouped, for example, when trying to select “Tanaka”, It is possible to prevent the stroke data belonging to “Yamamoto” from being selected at the same time. In addition, as shown in FIG. 13, grouping is performed not in a series of character strings but in units of characters, so that characters in different character strings can be simultaneously selected. Will be possible.
[0051]
Therefore, the stroke data determination unit 54 determines the recognition target category for each group of the stroke data. In order to determine the recognition target category on a group basis, for example, it is conceivable that the determination is made by majority decision of the individual determination results. When the determination results are the same (for example, when two strokes were determined to be a figure and a character in a group of four strokes, etc.), it is only necessary to determine that any of the recognition target categories is possible. However, it is not limited to these methods.
[0052]
Also, the stroke data selection unit 56 selects a group to be recognized in units of stroke data groups. For that purpose, when selecting a stroke, the center point of the circumscribed frame of each group may be obtained, and the stroke may be selected in group units depending on whether the center point is surrounded by the surrounding gesture. Of course, it is not limited to these methods.
[0053]
Various methods are also conceivable for grouping the stroke data. For example, as shown in FIG. 14, a method of dividing a set of stroke data for each line can be considered.
[0054]
First, as shown in FIG. 14A, it is conceivable to roughly divide a set of stroke data using the writing time. Specifically, a set of stroke data is arranged in the order of written time, and the time interval between adjacent stroke data (the time interval from the end point of the previous stroke data to the start point of the next stroke data) is 1 second. The vacant positions are set as division points. Hereinafter, a subset of the stroke data divided by the time interval is referred to as a “time group”.
[0055]
Next, attention is paid to the positional relationship of the stroke data in each time group. The locus of movement of the pen between stroke data is called an off-stroke. If the length of the off-stroke meets any of the following conditions, it can be determined that the off-stroke is a line break. The stroke data size means the length of the long side of the frame circumscribing the stroke data.
[0056]
(Condition 1) When the average value of the stroke data size in the time group is s, and the length of the off-stroke is longer than three times s
(Condition 2) When the writing direction is the horizontal direction and the average value of the stroke data size in the time group is s, the horizontal length of the off-stroke data is larger than s
For example, in FIG. 14B, the off-stroke is indicated by a thin broken line, and the off-stroke serving as a line break is indicated by a thick broken line. As shown in FIG. 14B, when writing the upper line and moving to the lower line, the length of the off-stroke becomes long, so it is presumed that a line break has occurred here. It becomes possible.
[0057]
FIG. 15 shows a configuration diagram of the group formation unit 111 in the case of using a time group. As shown in FIG. 15, the grouping unit 111 includes a stroke data time information extracting unit 151 and a stroke data group determining unit 152.
[0058]
The stroke data time information extraction unit 151 extracts a position where the time interval between adjacent stroke data (the time interval from the end point of the previous stroke data to the start point of the next stroke data) is 1 second or more, and extracts the stroke. The data group determination unit 152 specifies a delimiting portion when dividing in groups based on the off-stroke.
[0059]
At a position where a long off-stroke exists, the time interval is usually long. Therefore, even if the processing is not divided into two steps as shown in FIG. 15 and divided only by the time interval of the stroke, it is considered that it is practically sufficient.
[0060]
As another method of grouping stroke data, a method of grouping using a character recognition function can be considered. That is, this is a method of performing frameless character recognition for recognizing a character string written on a screen without using a character frame, and performing grouping based on the recognition result.
[0061]
FIG. 16 shows a configuration diagram of the grouping unit 111 in the case of using the result of frameless character recognition. As shown in FIG. 16, the group formation unit 111 includes a character recognition unit 161 and a stroke data group determination unit 162.
[0062]
More specifically, as shown in FIG. 17, the character recognition unit 161 performs frameless character recognition for recognizing a character string written on the screen without using a character frame, and divides the character into individual characters. A recognition result is obtained. Then, an evaluation value is added to each character recognition result as an output of the character recognition unit 161.
[0063]
Various evaluation values can be considered as the evaluation value output by the character recognition unit 161. For example, if the evaluation value is a perfect score of 1000 and the evaluation value exceeds 750, it is determined that the character recognition result is sufficiently reliable. In FIG. 17, each character recognition evaluation value is obtained, and if it is 750 or more, it is recognized as a character, and if it is 700 or less, it is recognized as a figure. If the score is between 700 points and 750 points, it can be determined that selection is possible at the time of character recognition and at the time of graphic recognition.
[0064]
The stroke data group determination unit 162 refers to the determination result as to whether or not the character is a character based on the evaluation value, and determines whether the recognition target category is a character, a graphic, or the like. Will be grouped.
[0065]
Next, a processing flow of a program for realizing the online handwritten information recognition apparatus according to the second embodiment of the present invention will be described. FIG. 18 shows a flowchart of processing of a program for realizing the online handwritten information recognition apparatus according to the embodiment of the present invention.
[0066]
In FIG. 18, first, the position coordinates of the pen are acquired, and coordinate point data is acquired including information on whether or not the stylus pen is being written (on state) (off state) (step S1801). Then, stroke data is extracted based on the acquired coordinate information, and stored as a set of stroke data (step S1802). Further, grouping of the stroke data is performed based on the time information of the stroke data (step S1803).
[0067]
Next, a recognition target category is set (step S1804), and it is determined whether or not the stored stroke data set group matches the recognition target category (step S1805).
[0068]
If the stroke data group matches the recognition target category (step S1805: YES), it is determined that only the stroke data included in the group is to be selected (step S1806), and the stroke data group is recognized. If it does not match the target category (Step S1805: No), it is determined that the stroke data included in the group is not a selection target (Step S1807). Then, the processing from step S1805 onward is repeated until the determination is completed for all the stroke data (step S1808: No).
[0069]
If the determination has been completed for all the stroke data (step S1808: YES), a group to be recognized is selected from the groups determined to be selected (step S1809). Finally, character recognition or graphic recognition is performed on the stroke data included in the selected group (step S1810).
[0070]
As described above, according to the second embodiment, the recognition target category is set before the start of the stroke data selection process, and the stroke data is grouped in advance, so that the recognition target category is included in a group having a different recognition target category. Stroke data that has been selected can be excluded from the recognition target, and even if the handwriting to be selected and the handwriting not to be selected exist close to each other, and the recognition target categories of both are the same, it can be easily performed. Only the desired handwriting can be selected.
[0071]
As shown in FIG. 19, the program for realizing the online handwritten information recognition device according to the embodiment of the present invention includes not only a portable recording medium 192 such as a CD-ROM 192-1 or a flexible disk 192-2, but also a program. The program may be stored in any of another storage device 191 provided at the end of the communication line, or a recording medium 194 such as a hard disk or a RAM of the computer 193. When the program is executed, the program is loaded and the main memory is stored. Run on
[0072]
Further, as shown in FIG. 19, the grouped stroke data and the like used by the online handwritten information recognition apparatus according to the embodiment of the present invention are also portable, such as a CD-ROM 192-1 and a flexible disk 192-2. Not only the recording medium 192, but also any other storage device 191 provided at the end of the communication line, or a recording medium 194 such as a hard disk or a RAM of the computer 193 may be used. It is read by the computer 193 when using such an online handwritten information recognition device.
[0073]
【The invention's effect】
As described above, according to the online handwritten information recognition device according to the present invention, when selecting and recognizing only a portion to be recognized from handwritten handwritten information in which various handwritings such as a character string and a figure are mixedly written, The target portion and the non-recognition target portion can be easily separated, and only the recognition target portion can be selected. As a result, the operability of handwriting selection is improved and selection errors are reduced, so that the usability of the handwriting recognition device is improved.
[Brief description of the drawings]
FIG. 1 is an explanatory diagram of processing in conventional online handwritten information recognition.
FIG. 2 is a configuration diagram of a conventional online handwritten information recognition device.
FIG. 3 is an explanatory diagram of a problem in a conventional online handwritten information recognition device.
FIG. 4 is an explanatory diagram of a problem in a conventional online handwritten information recognition device.
FIG. 5 is a configuration diagram of an online handwritten information recognition device according to the first embodiment of the present invention;
FIG. 6 is an exemplary data configuration of a stroke data storage unit in the online handwritten information recognition device according to the first embodiment of the present invention;
FIG. 7 is an explanatory diagram showing whether or not stroke data can be selected in the online handwritten information recognition apparatus according to the first embodiment of the present invention;
FIG. 8 is an explanatory diagram of stroke data selection by a gesture in the online handwritten information recognition apparatus according to the first embodiment of the present invention;
FIG. 9 is an explanatory diagram of stroke data selection by a gesture in the online handwritten information recognition apparatus according to the first embodiment of the present invention.
FIG. 10 is a flowchart of processing in the online handwritten information recognition device according to the first embodiment of the present invention;
FIG. 11 is a configuration diagram of an online handwritten information recognition apparatus according to a second embodiment of the present invention;
FIG. 12 is an exemplary diagram of grouping in the online handwritten information recognition apparatus according to the second embodiment of the present invention;
FIG. 13 is an exemplary diagram of grouping in the online handwritten information recognition apparatus according to the second embodiment of the present invention;
FIG. 14 is an explanatory diagram of a grouping method in the online handwritten information recognition apparatus according to the second embodiment of the present invention;
FIG. 15 is a configuration diagram of a group formation unit in the online handwritten information recognition apparatus according to the second embodiment of the present invention;
FIG. 16 is a configuration diagram of a group formation unit in the online handwritten information recognition device according to the second embodiment of the present invention;
FIG. 17 is an explanatory diagram of a grouping method using character recognition in the online handwritten information recognition apparatus according to the second embodiment of the present invention;
FIG. 18 is a flowchart of a process in the online handwritten information recognition apparatus according to the second embodiment of the present invention;
FIG. 19 is an exemplary diagram of a computer environment.
[Explanation of symbols]
21, 51 coordinate input unit
22, 52 Stroke data storage
23, 56 Stroke data selector
24, 57 Handwritten information recognition unit
53 Recognition target setting section
54 Stroke data judgment unit
55 Stroke data selection target limited section
111 Group Composition Department
151 Stroke data time information extraction unit
152, 162 stroke data group determination unit
161 character recognition unit
191 Line storage device
192 Portable recording media such as CD-ROM and flexible disk
192-1 CD-ROM
192-2 Flexible disk
193 Computer
194 Recording media such as RAM / hard disk on computer

Claims (8)

A coordinate input unit for inputting handwriting coordinate data,
For a set of stroke data formed by the input handwriting coordinate data, a handwriting information recognition unit that performs character recognition or figure recognition,
A stroke data selection unit that selects a subset of stroke data to be subjected to handwriting recognition from the set of stroke data,
An online handwritten information recognition device that performs character recognition or figure recognition in the handwritten information recognition unit on a subset of the selected stroke data, and outputs a recognition result,
A recognition target setting unit that sets a recognition target category to be recognized in the handwritten information recognition unit;
A stroke data determination unit that determines whether the stroke data belongs to the currently set recognition target category,
A stroke data selection target limiting unit that limits the stroke data to be selected by the stroke data selection unit to only the stroke data determined to belong to the recognition target category. apparatus. A group formation unit that classifies a plurality of stroke data for each group,
2. The online handwritten information recognition apparatus according to claim 1, wherein the stroke data determination unit determines the recognition target category to which the stroke data belongs in the group unit. A group formation unit that classifies a plurality of stroke data for each group,
The online handwritten information recognition apparatus according to claim 1, wherein the stroke data selection unit selects the stroke data for each group. 4. The online handwritten information recognition apparatus according to claim 2, wherein the group forming unit classifies the stroke data into the groups based on an order in which the stroke data is written. 5. 4. The online handwritten information recognition apparatus according to claim 2, wherein the group forming unit classifies the stroke data into groups based on time information including at least a time interval at which the stroke data is written. 5. The online handwritten information recognition apparatus according to any one of claims 1 to 3, wherein the group forming unit classifies the stroke data into the groups based on a result of character recognition of the stroke data. A process for inputting handwriting coordinate data;
A step of performing character recognition or figure recognition on a set of stroke data formed by the input handwriting coordinate data,
Selecting a subset of stroke data to be subjected to handwriting recognition from the set of stroke data,
An online handwritten information recognition method for performing character recognition or graphic recognition on a subset of the selected stroke data and outputting a recognition result,
Setting a recognition target category to be recognized;
A step of determining whether the stroke data belongs to the currently set recognition target category,
Limiting the stroke data to be selected to only the stroke data determined to belong to the recognition target category. Steps for inputting handwriting coordinate data;
Performing a character recognition or a figure recognition on a set of stroke data formed by the input handwriting coordinate data;
Selecting a subset of stroke data to be subjected to handwriting recognition from the set of stroke data,
A computer-executable program embodying an online handwritten information recognition method for performing character recognition or graphic recognition on a selected subset of the stroke data and outputting a recognition result,
Setting a recognition target category to be recognized;
Determining whether the stroke data belongs to the currently set recognition target category,
Limiting the stroke data to be selected to only the stroke data determined to belong to the recognition target category.

Images