JP2962911B2 - Character recognition device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a handwritten character recognition device having no restriction on the stroke order.

[0002]

2. Description of the Related Art In recent years, as an input device such as a word processor or the like, a handwritten character which is directly recognized on a tablet for character input or the like is recognized as a specific character and input, instead of a keyboard input requiring a certain level of skill. Character recognition devices have been developed and put into practical use.

Conventionally, in this type of handwritten character recognition device,
By classifying strokes of written characters into predetermined basic stroke codes and processing the code information from the pattern information, a reduction in dictionary capacity or an increase in processing speed is realized. However, this type of handwritten character recognition device has a restriction that the stroke order must be written correctly. However, the writing condition of writing all characters in the correct stroke order places a heavy burden on the scribe. In addition, it is rare for a writer to memorize all the correct stroke orders such as kanji with many strokes.

[0004] In order to solve the difficulties of such a conventional handwritten character recognition device, it is conceivable to preliminarily write, in a dictionary, locations where the stroke order is likely to fluctuate. (Details in Japanese Patent Publication No. 02-32673) However, in this method, it is necessary to preliminarily describe, in a character recognition dictionary, information on stroke order variation for all characters to be recognized. It is impractical to describe every stroke order variation for every character.

It is also conceivable that the correspondence between the strokes of the written characters and the standard characters is made for any variation in the stroke order. However, it is effective for characters having a small number of strokes, but in the case of kanji having a large number of strokes, The number of combinations of stroke order fluctuations becomes enormous, and the amount of calculation required for recognition increases rapidly, which poses a practical problem.

[0006]

The character recognition apparatus of the present invention has been made in consideration of the above-mentioned circumstances, and uses a recognition dictionary of a standard pattern in a correct stroke order without increasing the amount of calculation. An object of the present invention is to realize a character recognition device that recognizes written characters in the incorrect stroke order.

[0007]

According to the present invention, there is provided a character recognition apparatus comprising: a coordinate input unit for inputting handwriting coordinates; and a stroke input unit for forming a character based on time-series coordinate data obtained from the coordinate input unit. A stroke identification unit that classifies the shape into a predetermined basic stroke code, a basic stroke dictionary unit that describes the graphic characteristics of the basic stroke to identify the shape of the stroke, and a basic stroke that constitutes a character for each character to be processed And a character identification unit that identifies a character by comparing the basic stroke identified by the stroke identification unit with the character identification dictionary unit. A sentence based on a start / end point coordinate dictionary unit that describes the start and end points of the strokes constituting the character and time-series coordinate data obtained from the coordinate input unit. A stroke start / end point coordinate detecting unit for detecting start / end point coordinates of a stroke, and a stroke start / end point coordinate of a writing character obtained from the stroke start / end point coordinate detection unit and the start / end point coordinate dictionary unit. A stroke handling unit that handles strokes of characters,
A stroke rearranging unit for rearranging strokes of handwritten characters or standard characters based on the result information corresponding to the strokes, and a control unit for controlling them are further provided.

[0008]

According to the character recognition device of the present invention, a stroke of an input character is identified as a basic stroke code, and in a character recognition device for specifying a character, in order to recognize a character written in an erroneous stroke order, a character input device is used. Based on the obtained time-series coordinate data, the stroke correspondence unit compares the start and end point coordinates of the stroke detected by the stroke start and end point coordinate detection unit with the start and end point coordinate dictionary unit, and determines the stroke of the input character and the standard character. By responding and then rearranging the strokes by the stroke rearranging unit based on the result, it is possible to realize an online character recognition device with no restriction on the stroke order.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The details of the present invention will be described below with reference to embodiments. FIG. 1 is a functional block diagram of one embodiment of the character recognition device of the present invention. In the figure, a character pattern input unit 1
Consists of existing coordinate input devices such as tablets,
Time series coordinate data of input pattern (X coordinate, Y coordinate)
To the character pattern storage unit 2.

The character pattern storage unit 2 is composed of a RAM, and stores coordinate data input from the character pattern input unit 1. The stroke identification unit 3 includes a character pattern storage unit 2
ON / O of pen in time series coordinate data stored in
Separate into strokes based on FF information. These are separated into basic stroke shapes described in advance in the stroke identification dictionary unit 4 to obtain stroke codes.

The stroke identification dictionary unit 4 is composed of a ROM or a RAM, and describes graphic characteristics of a basic stroke for identifying a stroke shape. The character identification unit 5 specifies the input character by comparing the stroke code obtained by the stroke identification unit 3 with a basic stroke code string described in advance in the character identification dictionary unit 6.

The character identification dictionary unit 6 is a ROM or RA
A basic stroke code string consisting of M and constituting a character is described for each character to be processed. The detailed identification section 7 refers to a detailed identification dictionary 8 including a ROM or a RAM in which information such as a relative position of a stroke is described, and specifies a recognized character. The start / end point coordinate detection unit 9 detects the start / end point coordinates of each stroke from the input character pattern stored in the character pattern storage unit 2.

The stroke correspondence unit 11 includes start and end point coordinate information of each stroke of the input character obtained by the start and end point coordinate detection unit 9 and start and end of each stroke coordinate of the standard character stored in the standard character start and end point dictionary unit 12. Stroke correspondence is performed based on the point coordinate information. The standard character start / end point dictionary unit 12
It consists of a ROM or a RAM, and describes the starting point and the ending point of a stroke constituting a character for each character to be processed.

The stroke rearranging unit 10 rearranges the strokes of the input pattern stored in the character pattern storage unit 2 using the stroke correspondence information sent from the stroke correspondence unit 11. Recognition result output unit 13
Is composed of a display device such as a display or a printing device such as a printer, and outputs a recognition result.

The control unit 14 is mainly composed of a CPU and performs various controls such as control of the above-mentioned 2, 3, 5, 7, 9, 10, 11, and 13 and control of data between the units. Next, the operation of each unit will be described in detail by taking as an example a case of recognizing the kanji character "earth" shown in FIG. The handwriting coordinate data input from the character pattern input unit 1 is sent to the character pattern storage unit 2 and stored.

Next, the stroke identification unit 3 separates the time-series coordinate data recorded in the character pattern storage unit 2 into strokes based on the ON / OFF information of the pen, and describes these in the stroke identification dictionary 4 in advance. The stroke code is obtained by separating the stroke code into the basic stroke shape. For example, the first stroke of the kanji “earth” is “one”,
The stroke code assigned to each basic stroke is “|” for the second stroke, and “1” for the third stroke.

The control unit 14 sends the stroke code thus obtained to the character identification unit 5. The character identification unit 5 specifies the input character by comparing the stroke code sequence obtained from the input character with the stroke code sequence of each character described in the character identification dictionary unit 6. However, as in the case of the kanji “shi” for the kanji “earth” described above, there is a case where the stroke cannot be uniquely determined only by the shape of the stroke. And outputs the result to the output unit 13. Thereafter, the output unit 13 outputs the recognition result.

In the case of the conventional handwritten character recognition device, if the input character cannot be recognized despite the above-described processing, a result indicating that the input character cannot be recognized is output to the output unit 13. You. However, in the apparatus of the present invention, if the character cannot be recognized in spite of the basic character recognition processing described above, it is determined that the stroke order is different from the normal character input, and the strokes of the input character are rearranged. It is configured to perform processing and perform character recognition.

That is, the apparatus of the present invention has a configuration similar to that of the above-described conventional apparatus, and further includes a start / end point coordinate detection section 9, a stroke rearrangement section 10, a stroke correspondence section 11, and a start / end point coordinate detection section 12.
It has. Hereinafter, a process of recognizing a handwritten character in the erroneous stroke order which is a feature of the present invention will be described in detail.

The input character pattern stored in the character pattern storage unit 2 includes n strokes p1, p2,.
, Pn, and the j-th stroke pj is expressed by Expressions 1 and 2. (Pj1, pj2, pj3,..., Pjm (j)) Expression 1 where pji: the writing point of the jth stroke m (j): the total number of writing points of the jth stroke pji = (pjix, pjiy) Expression 2 The start and end point coordinates of each stroke are detected from the input character pattern stored in the character pattern storage section 2 by the start and end point coordinate detection section 9. Coordinate sequence sj of the starting point of each stroke
Are represented by Expressions 3 and 4, and the coordinate sequence ej of the end point is represented by Expressions 5 and 6.

(S1, s2, s3,..., Sn) Expression 3 where sj: the starting point coordinate of the j-th stroke n: the number of strokes sj = (pj1x, pj1y) Expression 4 (e1, e2, e3,...) , En) Equation 5 where ej: the end point coordinate of the j-th stroke n: the number of strokes ej = (pjm (j) x, pjm (j) y) Equation 6 Next, in the stroke corresponding unit 11, the start / end point coordinate detection unit 9 Strokes are associated with each other based on the start / end point coordinate information of each stroke of the input character and the start / end point coordinate information of each stroke coordinate of the standard character stored in the standard character start / end point dictionary unit 12.

At this time, the start / end point coordinate information of each stroke coordinate of the standard character stored in the standard character start / end point dictionary section 12 is obtained by converting the standard character pattern into n strokes P1, P
2,..., Pn, and the j-th stroke Pj is expressed by Equation 7,
If it can be expressed by Expression 8, the coordinate sequence Sj of the starting point of each stroke
Can be expressed by Expressions 9 and 10, and the coordinate sequence Ej of the end point can be expressed by Expressions 11 and 12.

[0023]

(Equation 1)

The inter-stroke distance d between the j-th stroke of the standard character pattern and the i-th stroke of the input character pattern
(I, j) can be expressed by Expression 13 by the city block distance between the start point and the end point. d (i, j) = | pj1x-Pi1x | + | pj1y-Pi1y | Expression 13 The distance between strokes is obtained for all the stroke pairs of the standard pattern and the input pattern by Expression (13). Stroke association is to determine the stroke of the input pattern corresponding to the i-th stroke of the standard pattern. If the stroke of the input pattern corresponding to the i-th stroke of the standard pattern is represented by K (i), Equation 14 is minimized, a stroke pair that satisfies Equation 15 is determined, and the stroke pair information is sent to the stroke rearranging unit 10. .

[0025]

(Equation 2)

The stroke rearranging unit 10 rearranges the strokes of the input pattern stored in the character pattern storage unit 2 using the stroke pair information sent from the stroke corresponding unit 11. The control unit 14 performs character recognition on the input pattern in which the strokes have been rearranged using the functions of the stroke identification unit 3, the character identification unit 5, and the detailed identification unit 7, and outputs the result to the output unit 13. .

[0027]

According to the character recognition device of the present invention, a character recognition device for identifying a stroke of an input character as a basic stroke code and specifying a character has a function of recognizing a character written in the wrong stroke order. Based on the point coordinates, the input characters correspond to the strokes of the standard characters, and by rearranging the strokes, an online character recognition device with no restriction on the stroke order can be realized.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a character recognition device according to an embodiment of the present invention.

FIG. 2 is a schematic diagram showing a stroke sequence in a correct stroke order of a kanji “earth”.

[Explanation of symbols]

 Reference Signs List 1 character pattern input unit 2 character pattern storage unit 3 stroke identification unit 4 stroke identification dictionary unit 5 character identification unit 6 character identification dictionary unit 7 detailed identification unit 8 detailed identification dictionary unit 9 start / end point coordinate detection unit 10 stroke rearrangement unit 11 stroke Corresponding unit 12 Standard character start / end point dictionary unit 13 Recognition result output unit 14 Control unit

Continuation of the front page (72) Inventor Hiroshi Horii 2--18 Keihanhondori, Moriguchi-shi, Osaka Sanyo Electric Co., Ltd. (56) References JP-A-61-295588 (JP, A) JP-A-59-188604 (JP, A) JP 5-62391 (JP, B2) (58) Fields investigated (Int. Cl. ⁶ , DB name) G06K 9/00-9/82

Claims (1)

(57) [Claims] 1. A coordinate input unit for inputting handwriting coordinates, and a stroke for classifying strokes constituting characters into a predetermined basic stroke code based on time-series coordinate data obtained from the coordinate input unit An identification unit;
A basic stroke dictionary unit that describes the graphic characteristics of the basic strokes for identifying the shape of the stroke, a character identification dictionary unit that describes the basic strokes that compose the character for each character to be processed, and the stroke identification unit A character recognition device comprising a character identification unit for identifying a character by comparing a basic stroke with the character identification dictionary unit, wherein a start / end point coordinate dictionary describing a start point and an end point of a stroke constituting a character for each processing target character Unit, a stroke start / end point coordinate detection unit for detecting a start / end point coordinate of a stroke constituting a character based on time-series coordinate data obtained from the coordinate input unit, and a handwritten character obtained from the stroke start / end point coordinate detection unit The start and end point coordinates of the stroke are compared with the start and end point coordinate dictionary unit, and the stroke of the written character is compared with the start and end point coordinate dictionary unit.
A stroke correspondence unit that performs stroke correspondence between input characters and standard characters based on the distance between the strokes according to the start and end points, and a stroke that rearranges strokes of written characters or standard characters based on the stroke correspondence result information. A character recognition device, comprising: a rearrangement unit; and a control unit that controls them.