CN101344918A - Method and device for judging stroke position normalization of handwritten Chinese characters - Google Patents

Abstract

The invention relates to a technology for judging the position normalization of strokes of handwritten Chinese characters, in particular to a method and a device for judging the position normalization of the strokes of the handwritten Chinese characters. The method for judging the stroke position normalization of the handwritten Chinese character comprises the steps of obtaining strokes of the handwritten Chinese character; acquiring a vector of the handwritten Chinese character pointing to the starting point of the (k-1) th stroke in the stroke order direction according to the starting point of the k stroke in the stroke order direction; acquiring a vector of a standard word stock Chinese character corresponding to the handwritten Chinese character, wherein the vector points to the starting point of the (k-1) th stroke in the stroke order direction according to the starting point of the k-th stroke in the stroke order direction; calculating an included angle theta between the stroke vector of the handwritten Chinese character and the stroke vector of the Chinese character in a standard word stock, and judging whether the included angle theta is within a preset error range delta theta or not, wherein if yes, the stroke position of the handwritten Chinese character is standard, and otherwise, the stroke position of the handwritten Chinese character is not standard; the technical scheme can identify the condition that the writing orders of the handwritten Chinese characters and the strokes of the same stroke shape are incorrect.

Description

The determination methods of stroke position standard of handwritten Chinese character and device

Technical field

The present invention relates to the judgment technology of stroke position standard of handwritten Chinese character, relate in particular to a kind of determination methods and device of stroke position standard of handwritten Chinese character.

Background technology

Existing handwriting recognition technology is in user writing, coordinate by hand-written direction (time order and function order) record picture element, and the vector value of hand-written direction, by stroke (being called stroke again) it is divided into different arrangement sets, these records is handled each the picture element standardization that makes in each arrangement set; Then, the standard letter arrangement set in hand-written script arrangement set and the standard character library is compared, get the highest Chinese character of similarity, return to the user as recognition result.

From above prior art as can be seen, it is with the hand-written arrangement set similarity of the whole handwritten Chinese character standard as the identification Chinese character, such as Chinese character " two ", the correct order of strokes observed in calligraphy should be to write top horizontal picture earlier, write following horizontal picture again, write following horizontal picture earlier, write top horizontal picture again as the user, prior art is still thought and is write correctly, and can't judge the stroke position standardization of handwritten Chinese character.

Summary of the invention

The invention provides a kind of determination methods and device of stroke position standard of handwritten Chinese character, it is by the angle of handwritten Chinese character two-stroke starting point vector with standard character library Chinese character two-stroke starting point vector, compare with the error range that presets, thereby judge the stroke position standardization.

A kind of determination methods of stroke position standard of handwritten Chinese character comprises: A) obtain the stroke of handwritten Chinese character, calculate the stroke number n of handwritten Chinese character; Default k=2; B) obtain handwritten Chinese character according to stroke order the starting point of the k stroke of direction point to the vector of the starting point of (k-1) stroke of direction according to stroke order; C) the standard character library Chinese character that obtains the handwritten Chinese character correspondence according to stroke order the starting point of the k stroke of direction point to the vector of the starting point of (k-1) stroke of direction according to stroke order; D) calculate the angle theta of hand-written Chinese character strokes vector and standard character library Chinese-character stroke vector, judge angle theta whether in the error range Δ θ that presets, if, the stroke position standard of handwritten Chinese character, otherwise the stroke position of handwritten Chinese character is lack of standardization; E) judge whether (k+1)＞n sets up, if the judgement of stroke position standard of handwritten Chinese character finishes, otherwise k=(k+1) continues step B.

Wherein, the font of standard character library Chinese character is " regular script-GB2312 ", and size is got 248 pounds of font sizes; Angle theta is specially:

$\mathrm{\&theta;}=\left\{\begin{array}{cc}\mathrm{\&theta;}& 0\&le;\mathrm{\&theta;}\&le;\mathrm{\&pi;}\\ 2\mathrm{\&pi;}-\mathrm{\&theta;}& \mathrm{\&pi;}<\mathrm{\&theta;}<2\mathrm{\&pi;}\end{array}\right.$

The error range Δ θ that presets is specially:

Wherein, len represents the standard character library Chinese character distance of the starting point of the starting point of the k stroke of direction and (k-1) stroke according to stroke order.

A kind of judgment means of stroke position standard of handwritten Chinese character comprises: the hand-written Chinese character strokes acquiring unit, be used to obtain the stroke of handwritten Chinese character, and calculate the stroke number n of handwritten Chinese character; Parameter is preset the unit, is used for default k=2; Handwritten Chinese character vector acquiring unit, be used to obtain handwritten Chinese character according to stroke order the starting point of the k stroke of direction point to the vector of the starting point of (k-1) stroke of direction according to stroke order; Standard Chinese character vector acquiring unit, the standard character library Chinese character that is used to the to obtain the handwritten Chinese character correspondence according to stroke order starting point of the k stroke of direction points to the vector of the starting point of (k-1) stroke of direction according to stroke order; The angle computing unit is used to calculate the angle theta of the standard character library Chinese-character stroke vector that hand-written Chinese character strokes vector that handwritten Chinese character vector acquiring unit obtains and standard Chinese character vector acquiring unit obtain; The standardization judging unit is used to judge angle theta whether in the error range Δ θ that presets, if, the stroke position standard of handwritten Chinese character, otherwise the stroke position of handwritten Chinese character is lack of standardization; The cycle criterion unit is used for judging whether (k+1)＞n sets up, if the judgement of stroke position standard of handwritten Chinese character finishes, otherwise k=(k+1) sends the request of obtaining vector to handwritten Chinese character vector acquiring unit and standard Chinese character vector acquiring unit.

Wherein, the font of standard character library Chinese character is " regular script-GB2312 ", and size is got 248 pounds of font sizes; The angle theta of angle computing unit is specially:

$\mathrm{\&theta;}=\left\{\begin{array}{cc}\mathrm{\&theta;}& 0\&le;\mathrm{\&theta;}\&le;\mathrm{\&pi;}\\ 2\mathrm{\&pi;}-\mathrm{\&theta;}& \mathrm{\&pi;}<\mathrm{\&theta;}<2\mathrm{\&pi;}\end{array}\right.$

The error range Δ θ of standardization judging unit is specially:

Wherein, len represents the standard character library Chinese character distance of the starting point of the starting point of the k stroke of direction and (k-1) stroke according to stroke order.

As can be seen from the above technical solutions, in the present invention, A) obtain the stroke of handwritten Chinese character, calculate the stroke number n of handwritten Chinese character; Default k=2; B) obtain handwritten Chinese character according to stroke order the starting point of the k stroke of direction point to the vector of the starting point of (k-1) stroke of direction according to stroke order; C) the standard character library Chinese character that obtains the handwritten Chinese character correspondence according to stroke order the starting point of the k stroke of direction point to the vector of the starting point of (k-1) stroke of direction according to stroke order; D) calculate the angle theta of hand-written Chinese character strokes vector and standard character library Chinese-character stroke vector, judge angle theta whether in the error range Δ θ that presets, if, the stroke position standard of handwritten Chinese character, otherwise the stroke position of handwritten Chinese character is lack of standardization; E) judge whether (k+1)＞n sets up, if the judgement of stroke position standard of handwritten Chinese character finishes, otherwise k=(k+1) continues step B.The technical program compares with the error range that presets, thereby judges the stroke position standardization by the angle of handwritten Chinese character two-stroke starting point vector with standard character library Chinese character two-stroke starting point vector; The technical program can identify handwritten Chinese character with the incorrect situation of form of a stroke or a combination of strokes stroke order of writing strokes.

Description of drawings

Fig. 1 is the determination methods process flow diagram of the stroke position standard of handwritten Chinese character of the embodiment of the invention;

Fig. 2 is the synoptic diagram of the standard character library Chinese character " early " of the embodiment of the invention;

Fig. 3 is the synoptic diagram of the handwritten Chinese character " early " of the embodiment of the invention;

Fig. 4 is the judgment means synoptic diagram of the stroke position standard of handwritten Chinese character of the embodiment of the invention.

Embodiment

Embodiment one

Please refer to Fig. 1, be the determination methods process flow diagram of the stroke position standard of handwritten Chinese character of the embodiment of the invention.The determination methods of the stroke position standard of handwritten Chinese character of present embodiment may further comprise the steps.

101. obtain the stroke of handwritten Chinese character, calculate the stroke number n of handwritten Chinese character; Default k=2.

The above-mentioned method of obtaining hand-written Chinese character strokes can for, the user writes on touch-screen or handwriting pad, touch-screen or handwriting pad are converted to picture element with touch pressure, as hand-written Chinese character strokes, also can directly obtain stroke writing to this picture element with digital pen; Also can scan the font on this blank sheet of paper on blank sheet of paper, be converted to picture element for the user writes, this picture element as hand-written Chinese character strokes.In the present embodiment, can set up rectangular coordinate system, each picture element that obtains all has the coordinate of oneself.

102. obtain handwritten Chinese character according to stroke order the starting point of the k stroke of direction point to the vector of the starting point of (k-1) stroke of direction according to stroke order.

What the stroke position standardization was meant hand-written Chinese character strokes writes sequencing compliant, i.e. order of strokes observed in calligraphy compliant.When moving this step for the first time, k=2, the starting point of obtaining the 2nd stroke is pointed to the vector of the starting point of the 1st stroke.For the special Chinese character that has only a stroke, there is not the problem of writing sequencing of judging stroke, so not within the scope that the technical program is discussed.

103. the standard character library Chinese character that the obtains the handwritten Chinese character correspondence according to stroke order starting point of the k stroke of direction points to the vector of the starting point of (k-1) stroke of direction according to stroke order.

The standard character library Chinese character according to stroke order starting point of the k stroke of direction points to according to stroke order that the vector of the starting point of (k-1) stroke of direction can calculate in advance and store, and the vector that is about between the stroke of whole standard Chinese characters is preserved; Before calculating angle theta, directly read vector between the stroke of standard character library Chinese character of handwritten Chinese character correspondence.Also can from the standard character library, obtain the Chinese character of handwritten Chinese character correspondence, obtain the starting point coordinate of the k stroke of direction according to stroke order and point to the starting point coordinate of (k-1) stroke of direction according to stroke order, calculate their vector then.

104. calculate the angle theta of hand-written Chinese character strokes vector and standard character library Chinese-character stroke vector.

In the present embodiment, owing to set up rectangular coordinate system, so can calculate the angle theta of these two vectors easily.

105. judge that angle theta is whether in the error range Δ θ that presets.If, the stroke position standard of handwritten Chinese character, otherwise the stroke position of handwritten Chinese character is lack of standardization.

More particularly, this step is carried out the judgement of location specification to k stroke and (k-1) stroke of handwritten Chinese character.

106. judge whether (k+1)＞n sets up, if the judgement of stroke position standard of handwritten Chinese character finishes, otherwise k=(k+1) continues step 102.

The technical program is carried out the judgement of location specification to a plurality of strokes of a Chinese character.Such as a Chinese character 3 strokes are arranged, in operation for the first time, k=2, k+1=3, (k+1)＞n is false, so need continue operating procedure 102.

In the present embodiment, the font of standard character library Chinese character is " regular script-GB2312 ", and size is got 248 pounds of font sizes.

Above-mentioned angle theta is specially:

$\mathrm{\&theta;}=\left\{\begin{array}{cc}\mathrm{\&theta;}& 0\&le;\mathrm{\&theta;}\&le;\mathrm{\&pi;}\\ 2\mathrm{\&pi;}-\mathrm{\&theta;}& \mathrm{\&pi;}<\mathrm{\&theta;}<2\mathrm{\&pi;}\end{array}\right.$

The above-mentioned error range Δ θ that presets is specially:

Wherein, len represents the standard character library Chinese character distance of the starting point of the starting point of the k stroke of direction and (k-1) stroke according to stroke order.2 distances are divided into 7 sections, the error amount difference of each section permission, distance is near more, and the error amount of permission is big more.Need to prove that in actual applications, can determine 2 numbers apart from segmentation according to accuracy requirement, accuracy requirement is high more, segmentation is many more.

The font of present embodiment standard character library Chinese character is " regular script-GB2312 ", and size is got 248 pounds of font sizes, and is corresponding, handwritten Chinese character also will be normalized to 248 pounds, what is called is normalized to, and handwritten Chinese character is carried out on the whole convergent-divergent, makes its long and wide standard that meets 248 pounds.Need to prove, in actual applications, also can get other big small type size, its error range Δ θ can adjust according to the distance of segmentation number and every section.

Please refer to Fig. 2 and Fig. 3; Fig. 2 is the synoptic diagram of the standard character library Chinese character " early " of the embodiment of the invention; Fig. 3 is the synoptic diagram of the handwritten Chinese character " early " of the embodiment of the invention.The process of judging the stroke position standardization is as follows:

1. obtain the 2nd starting point and the 1st starting point vector of standard character library Chinese character " early ", its angle value is 4.712 radians, plays dot spacing from 2.000 pounds, obtains error range π; Obtain that vectorial angle value is 1.418 radians between the 2nd of handwritten Chinese character " early " and the 1st starting point, angle (2 π-the 3.294)＜π of hand-written Chinese character strokes vector and standard character library Chinese-character stroke vector is in the permissible error scope.

2. obtain the 3rd starting point and the 2nd starting point vector of standard character library Chinese character " early ", its angle value is 5.155 radians, plays dot spacing from 42.047 pounds, obtains error range 7 π/36; Obtain that vectorial angle value is 5.121 radians between the 3rd of handwritten Chinese character " early " and the 2nd starting point, angle 0.034＜7 π/36 of hand-written Chinese character strokes vector and standard character library Chinese-character stroke vector are in the permissible error scope.

3. obtain the 4th starting point and the 3rd starting point vector of standard character library Chinese character " early ", its angle value is 4.746 radians, plays dot spacing from 30.017 pounds, obtains error range 7 π/36; Obtain that vectorial angle value is 4.424 radians between the 4th of handwritten Chinese character " early " and the 3rd starting point, angle 0.322＜7 π/36 of hand-written Chinese character strokes vector and standard character library Chinese-character stroke vector are in the permissible error scope.

4. obtain the 5th starting point and the 4th starting point vector of standard character library Chinese character " early ", its angle value is 3.796 radians, plays dot spacing from 75.604 pounds, obtains error range π/6; Obtain that vectorial angle value is 3.726 radians between the 5th of handwritten Chinese character " early " and the 4th starting point, angle 0.070＜π/6 of hand-written Chinese character strokes vector and standard character library Chinese-character stroke vector are in the permissible error scope.

5. obtain the 6th starting point and the 5th starting point vector of standard character library Chinese character " early ", its angle value is 0.428 radian, plays dot spacing from 101.134 pounds, obtains error range 5 π/36; Obtain that vectorial angle value is 0.345 radian between the 6th of handwritten Chinese character " early " and the 5th starting point, angle 0.083＜5 π/36 of hand-written Chinese character strokes vector and standard character library Chinese-character stroke vector are in the permissible error scope.

Embodiment two

Please refer to Fig. 4, be the judgment means synoptic diagram of the stroke position standard of handwritten Chinese character of the embodiment of the invention.The judgment means of the stroke position standard of handwritten Chinese character of present embodiment comprises:

Hand-written Chinese character strokes acquiring unit 202 is used to obtain the stroke of handwritten Chinese character, calculates the stroke number n of handwritten Chinese character;

Parameter is preset unit 201, is used for default k=2;

Handwritten Chinese character vector acquiring unit 203, be used to obtain handwritten Chinese character according to stroke order the starting point of the k stroke of direction point to the vector of the starting point of (k-1) stroke of direction according to stroke order;

Standard Chinese character vector acquiring unit 204, the standard character library Chinese character that is used to the to obtain the handwritten Chinese character correspondence according to stroke order starting point of the k stroke of direction points to the vector of the starting point of (k-1) stroke of direction according to stroke order;

Angle computing unit 205 is used to calculate the angle theta of the standard character library Chinese-character stroke vector that hand-written Chinese character strokes vector that handwritten Chinese character vector acquiring unit 203 obtains and standard Chinese character vector acquiring unit 204 obtain;

Standardization judging unit 206 is used to judge angle theta whether in the error range Δ θ that presets, if, the stroke position standard of handwritten Chinese character, otherwise the stroke position of handwritten Chinese character is lack of standardization;

Cycle criterion unit 207 is used for judging whether (k+1)＞n sets up, if, the judgement of stroke position standard of handwritten Chinese character finishes, otherwise k=(k+1) sends the request of obtaining vector to handwritten Chinese character vector acquiring unit 203 and standard Chinese character vector acquiring unit 204.

Wherein, in the present embodiment, the font of standard character library Chinese character is " regular script-GB2312 ", and size is got 248 pounds of font sizes.

The angle theta of above-mentioned angle computing unit 205 is specially:

$\mathrm{\&theta;}=\left\{\begin{array}{cc}\mathrm{\&theta;}& 0\&le;\mathrm{\&theta;}\&le;\mathrm{\&pi;}\\ 2\mathrm{\&pi;}-\mathrm{\&theta;}& \mathrm{\&pi;}<\mathrm{\&theta;}<2\mathrm{\&pi;}\end{array}\right.$

The error range Δ θ of above-mentioned standardization judging unit 206 is specially:

Wherein, len represents the standard character library Chinese character distance of the starting point of the starting point of the k stroke of direction and (k-1) stroke according to stroke order.

Above content is preferred embodiment of the present invention only, and for those of ordinary skill in the art, according to thought of the present invention, the part that all can change in specific embodiments and applications, this description should not be construed as limitation of the present invention.

Claims (4)

1. the determination methods of a stroke position standard of handwritten Chinese character is characterized in that, comprising:

A) obtain the stroke of handwritten Chinese character, calculate the stroke number n of described handwritten Chinese character; Default k=2;

B) obtain described handwritten Chinese character according to stroke order the starting point of the k stroke of direction point to the vector of the starting point of (k-1) stroke of direction according to stroke order;

C) the standard character library Chinese character that obtains described handwritten Chinese character correspondence according to stroke order the starting point of the k stroke of direction point to the vector of the starting point of (k-1) stroke of direction according to stroke order;

D) angle theta of described hand-written Chinese character strokes vector of calculating and standard character library Chinese-character stroke vector, judge angle theta whether in the error range Δ θ that presets, if, the stroke position standard of described handwritten Chinese character, otherwise the stroke position of described handwritten Chinese character is lack of standardization;

E) judge whether (k+1)＞n sets up, if the judgement of stroke position standard of handwritten Chinese character finishes, otherwise k=(k+1) continues step B.

2. method according to claim 1 is characterized in that, the font of described standard character library Chinese character is " regular script-GB2312 ", and size is got 248 pounds of font sizes;

Described angle theta is specially:

$\mathrm{\&theta;}=\left\{\begin{array}{cc}\mathrm{\&theta;}& 0\&le;\mathrm{\&theta;}\&le;\mathrm{\&pi;}\\ 2\mathrm{\&pi;}-\mathrm{\&theta;}& \mathrm{\&pi;}<\mathrm{\&theta;}<2\mathrm{\&pi;}\end{array}\right.$

The described error range Δ θ that presets is specially:

Wherein, len represents the standard character library Chinese character distance of the starting point of the starting point of the k stroke of direction and (k-1) stroke according to stroke order.

3. the judgment means of a stroke position standard of handwritten Chinese character is characterized in that, comprising:

The hand-written Chinese character strokes acquiring unit is used to obtain the stroke of handwritten Chinese character, calculates the stroke number n of described handwritten Chinese character;

Parameter is preset the unit, is used for default k=2;

Handwritten Chinese character vector acquiring unit, be used to obtain described handwritten Chinese character according to stroke order the starting point of the k stroke of direction point to the vector of the starting point of (k-1) stroke of direction according to stroke order;

Standard Chinese character vector acquiring unit, the standard character library Chinese character that is used to the to obtain described handwritten Chinese character correspondence according to stroke order starting point of the k stroke of direction points to the vector of the starting point of (k-1) stroke of direction according to stroke order;

The angle computing unit is used to calculate the angle theta of the standard character library Chinese-character stroke vector that hand-written Chinese character strokes vector that described handwritten Chinese character vector acquiring unit obtains and described standard Chinese character vector acquiring unit obtain;

The standardization judging unit is used to judge angle theta whether in the error range Δ θ that presets, if, the stroke position standard of described handwritten Chinese character, otherwise the stroke position of described handwritten Chinese character is lack of standardization;

The cycle criterion unit is used for judging whether (k+1)＞n sets up, if the judgement of stroke position standard of handwritten Chinese character finishes, otherwise k=(k+1) sends the request of obtaining vector to handwritten Chinese character vector acquiring unit and standard Chinese character vector acquiring unit.

4. device according to claim 3 is characterized in that, the font of described standard character library Chinese character is " regular script-GB2312 ", and size is got 248 pounds of font sizes;

The angle theta of described angle computing unit is specially:

$\mathrm{\&theta;}=\left\{\begin{array}{cc}\mathrm{\&theta;}& 0\&le;\mathrm{\&theta;}\&le;\mathrm{\&pi;}\\ 2\mathrm{\&pi;}-\mathrm{\&theta;}& \mathrm{\&pi;}<\mathrm{\&theta;}<2\mathrm{\&pi;}\end{array}\right.$

The error range Δ θ of described standardization judging unit is specially:

Wherein, len represents the standard character library Chinese character distance of the starting point of the starting point of the k stroke of direction and (k-1) stroke according to stroke order.

Images