Identifying method of hand-written Latin letter

• US 20100135576A1
• Filed: 03/03/2009
• Published: 06/03/2010
• Est. Priority Date: 12/01/2008
• Status: Active Grant

• Current Assignee: Ningbo Sunrun ELEC & Info St & D Company Limited
• Original Assignee: Ningbo Sunrun ELEC & Info St & D Company Limited
• Inventors: Zhang, Qingbo, Chen, Jing, Chen, Ping, Shi, Zhihui, Ma, Chengchen, Ding, Hongzhen, Jia, Dexiang, He, Jiaming, Shi, Aijun, Wen, Jianfen, Fan, Zhouyi, Fan, Linghui
• Granted Patent: US 8,150,159 B2
• Time in Patent Office: Days
• US Class Current: 382/186
• CPC Class Codes: G06V 30/347 Sampling; Contour coding; Stroke extraction
• CPC Class Codes: G06V 30/36 Matching; Classification

Abstract

The present invention discloses an identifying method of hand-written Latin letter. The present invention considers many hand-written styles of Latin letter, extract many stable characteristics of Latin letter of different hand-written styles, and classify the Latin letter aggregation each time with one characteristic, so that the whole standard Latin letter aggregation is classified into many small Latin letter aggregations with intersection to be the coarse classification candidate letter aggregations to be identified. When identifying the inputted hand-written Latin letter, obtain the coarse classification candidate letter aggregation that matches with the characteristics of the inputted hand-written Latin letter. Many stable characteristics ensure the identifying rate. The multilayer coarse classification candidate letter aggregations regulate the searching path and increase the identifying speed.

Claims

1. An identifying method of hand-written Latin letter, comprising steps of:

(a) collecting chirography coordinates of hand-written Latin letter that are inputted in a terminal unit, and storing said chirography coordinates into a predefined structural array in real time;

(b) obtaining new chirography coordinate lattice by preprocessing chirography coordinates and chirography coordinates lattice in a matrix format transformed from said chirography coordinates stored in a structural array;

(c) Classifying all standard Latin letters in different layers according to characteristics of each Latin letter during the hand writting, to obtain a coarse classification candidate letter aggregation matching with the inputted hand-written Latin letter, wherein step (c) comprises steps of;

(c.1) classifying all standard Latin letter into up-down structure coarse classification candidate letter aggregation and non-up-down structure coarse classification candidate letter aggregation, according to whether the hand-written Latin letter is up-down structure or not, wherein the up-down structure coarse classification candidate letter aggregation includes standard Latin letters with inherent up-down structure during the hand-writing and standard Latin letters with up-down structure caused by the hand-writing, and the non-up-down structure coarse classification candidate letter aggregation includes the standard Latin letters without inherent up-down structure;

(c.2) classifying the up-down structure coarse classification candidate letter aggregation into transverse stroke coarse classification candidate letter aggregation and non-transverse stroke coarse classification candidate letter aggregation, according to whether the last stroke of the hand-written standard Latin letter is transverse stroke or not, and take these two coarse classification candidate letter aggregations as the coarse classification candidate letter aggregation that matches with the inputted hand-written Latin letter, wherein the transverse stroke coarse classification candidate letter aggregation includes the hand-written standard Latin letters with inherent transverse last stroke and the hand-written standard Latin letters with transverse last stroke caused by the hand writing;

the non-transverse stroke coarse classification candidate letter aggregation includes the hand-written standard Latin letters without inherent transverse last stroke;

classifying the non-up-down structure coarse classification candidate letter aggregation into the coarse classification candidate letter aggregations with H=1, 2, 3, and 4, according to the transverse cross number H during the hand writing;

(c.3) classifying the coarse classification candidate letter aggregation with H=1 into the coarse classification candidate letter aggregations with S=1, 2, 3, and 4, according to the vertical cross number S during the hand writing, and take these four coarse classification candidate letter aggregations as the coarse classification candidate letter aggregation that matches with the inputted hand-written Latin letter;

classifying the coarse classification candidate letter aggregation with H=2 into the coarse classification candidate letter aggregations with UP=1, 2, and 3, according to the transverse cross number UP in the ⅓

upper portion of the Latin letter during the hand writing, and take the UP=2 and UP=3 coarse classification candidate letter aggregations as the coarse classification candidate letter aggregation that matches with the inputted hand-written Latin letter;

classifying the coarse classification candidate letter aggregation with H=3 into the coarse classification candidate letter aggregations with Maxs=1, 2, and 3, according to the optimal vertical cross number Maxs during the hand writing, and take the Maxs=2 and Maxs=3 coarse classification candidate letter aggregations as the coarse classification candidate letter aggregation that matches with the inputted hand-written Latin letter;

classifying the coarse classification candidate letter aggregation with H=4 into the coarse classification candidate letter aggregations with Down=1, 2, and 3, according to the transverse cross number Down in the lower ⅓

portion of the Latin letter during the hand writing, and take the Down=2 and Down=3 coarse classification candidate letter aggregations as the coarse classification candidate letter aggregation that matches with the inputted hand-written Latin letter;

(c.4) classifying the coarse classification candidate letter aggregation with UP=1 into the coarse classification candidate letter aggregations with S1=1, 2, and 3, according to the vertical cross number S1 during the hand writing;

classifying the coarse classification candidate letter aggregation with Maxs=1 into the coarse classification candidate letter aggregations with first right-up position, first left-down position, first left-down position and first right-down position, according to the position information of the last new chirography coordinate point of the new chirography coordinate lattice in the standard Latin letter during the hand writing;

classifying the coarse classification candidate letter aggregation with Down=1 into the coarse classification candidate letter aggregations with S2=1, 2, and 3, according to the vertical cross number S2 during the hand writing;

(c.5) classifying the coarse classification candidate letter aggregation with S1=1 into the coarse classification candidate letter aggregations with second right-up position, second left-down position, second left-down position and second right-down position, according to the position information of the last new chirography coordinate point of the new chirography coordinate lattice in the standard Latin letter during the hand writing;

classifying the coarse classification candidate letter aggregation with S1=2 into the coarse classification candidate letter aggregations with third right-up position, third left-down position, third left-down position and third right-down position, according to the position information of the last new chirography coordinate point of the new chirography coordinate lattice in the standard Latin letter during the hand writing;

classifying the coarse classification candidate letter aggregation with S1=3 into the coarse classification candidate letter aggregations with fourth right-up position, fourth left-down position, fourth left-down position and fourth right-down position, according to the position information of the last new chirography coordinate point of the new chirography coordinate lattice in the standard Latin letter during the hand writing;

classifying the coarse classification candidate letter aggregation with first left-up position into the coarse classification candidate letter aggregations with first left arch opening and first non-left arch opening, according to whether there is left arch opening during the hand writing, and take the two coarse classification candidate letter aggregations as the coarse classification candidate letter aggregation that matches with the inputted hand-written Latin letter;

classifying the coarse classification candidate letter aggregation with first left-down position, first right-up position and first right-down into the coarse classification candidate letter aggregations with second left arch opening and second non-left arch opening, third left arch opening and third non-left arch opening, fourth left arch opening and fourth non-left arch opening respectively, according to whether there is left arch opening during the hand writing;

classifying the coarse classification candidate letter aggregation with S2=1 into the coarse classification candidate letter aggregations with fifth right-up position, fifth left-down position, fifth left-down position and fifth right-down position, according to the position information of the last new chirography coordinate point of the new chirography coordinate lattice in the standard Latin letter during the hand writing;

classifying the coarse classification candidate letter aggregation with S2=2 into the coarse classification candidate letter aggregations with sixth right-up position, sixth left-down position, sixth left-down position and sixth right-down position, according to the position information of the last new chirography coordinate point of the new chirography coordinate lattice in the standard Latin letter during the hand writing;

classifying the coarse classification candidate letter aggregation with S2=3 into the coarse classification candidate letter aggregations with seventh right-up position, seventh left-down position, seventh left-down position and seventh right-down position, according to the position information of the last new chirography coordinate point of the new chirography coordinate lattice in the standard Latin letter during the hand writing;

(c.6) classifying the coarse classification candidate letter aggregation with second left-up position into the coarse classification candidate letter aggregations with fifth left arch opening and fifth non-left arch opening, according whether there is left arch opening or not during the hand writing, and take the two coarse classification candidate letter aggregations as the coarse classification candidate letter aggregation that matches with the inputted hand-written Latin letter;

classifying the coarse classification candidate letter aggregation with second left-down position, second right-up position and second right-down into the coarse classification candidate letter aggregations with sixth left arch opening and sixth non-left arch opening, seventh left arch opening and seventh non-left arch opening, eighth left arch opening and eighth non-left arch opening respectively, according to whether there is left arch opening during the hand writing;

classifying the coarse classification candidate letter aggregation with third left-up position into the coarse classification candidate letter aggregations with ninth left arch opening and ninth non-left arch opening, according whether there is left arch opening or not during the hand writing, and take the two coarse classification candidate letter aggregations as the coarse classification candidate letter aggregation that matches with the inputted hand-written Latin letter;

classifying the coarse classification candidate letter aggregation with third left-down position, third right-up position and third right-down into the coarse classification candidate letter aggregations with tenth left arch opening and tenth non-left arch opening, eleventh left arch opening and eleventh non-left arch opening, twelfth left arch opening and twelfth non-left arch opening respectively, according to whether there is left arch opening during the hand writing;

classifying the coarse classification candidate letter aggregation with fourth left-up position into the coarse classification candidate letter aggregations with thirteenth left arch opening and thirteenth non-left arch opening, according whether there is left arch opening or not during the hand writing, and take the two coarse classification candidate letter aggregations as the coarse classification candidate letter aggregation that matches with the inputted hand-written Latin letter;

classifying the coarse classification candidate letter aggregation with fourth left-down position, fourth right-up position and fourth right-down into the coarse classification candidate letter aggregations with fourteenth left arch opening and fourteenth non-left arch opening, fifteenth left arch opening and fifteenth non-left arch opening, sixteenth left arch opening and sixteenth non-left arch opening respectively, according to whether there is left arch opening during the hand writing;

classifying the coarse classification candidate letter aggregation with fifth left-up position into the coarse classification candidate letter aggregations with seventeenth left arch opening and seventeenth non-left arch opening, according whether there is left arch opening or not during the hand writing, and take the two coarse classification candidate letter aggregations as the coarse classification candidate letter aggregation that matches with the inputted hand-written Latin letter;

classifying the coarse classification candidate letter aggregation with fifth left-down position, fifth right-up position and fifth right-down into the coarse classification candidate letter aggregations with eighteenth left arch opening and eighteenth non-left arch opening, nineteenth left arch opening and nineteenth non-left arch opening, twentieth left arch opening and twentieth non-left arch opening respectively, according to whether there is left arch opening during the hand writing;

classifying the coarse classification candidate letter aggregation with sixth left-up position into the coarse classification candidate letter aggregations with twenty-first left arch opening and twenty-first non-left arch opening, according whether there is left arch opening or not during the hand writing, and take the two coarse classification candidate letter aggregations as the coarse classification candidate letter aggregation that matches with the inputted hand-written Latin letter;

classifying the coarse classification candidate letter aggregation with sixth left-down position, sixth right-up position and sixth right-down into the coarse classification candidate letter aggregations with twenty-second left arch opening and twenty-second non-left arch opening, twenty-third left arch opening and twenty-third non-left arch opening, twenty-fourth left arch opening and twenty-fourth non-left arch opening respectively, according to whether there is left arch opening during the hand writing;

classifying the coarse classification candidate letter aggregation with seventh left-up position into the coarse classification candidate letter aggregations with twenty-fifth left arch opening and twenty-fifth non-left arch opening, according whether there is left arch opening or not during the hand writing, and take the two coarse classification candidate letter aggregations as the coarse classification candidate letter aggregation that matches with the inputted hand-written Latin letter;

classifying the coarse classification candidate letter aggregation with seventh left-down position, seventh right-up position and seventh right-down into the coarse classification candidate letter aggregations with twenty-sixth left arch opening and twenty-sixth non-left arch opening, twenty-seventh left arch opening and twenty-seventh non-left arch opening, twenty-eighth left arch opening and twenty-eighth non-left arch opening respectively, according to whether there is left arch opening during the hand writing;

(d) according to the local characteristic of the inputted hand-written Latin letter, matching the inputted hand-written Latin letter with the coarse classification candidate letter aggregation of each aggregation obtained in step (c), and obtain the coarse classification candidate letter aggregation matched with the inputted hand-written Latin letter;

(e) extracting a freeman chain code of inputted hand-written Latin letter;

calculating a matching probability between the freeman chain code of the inputted hand-written Latin letter and the optimal Hidden Markov Model of each standard letter stored in a predetermined letter library and corresponding to each letter stored in the coarse classification candidate letter aggregation matched with the inputted hand-written Latin letter;

obtaining an optimized matching probability from each matching probability;

determining that the standard letter stored in the predetermined letter library and corresponding to the optimized matching probability as the final identifying result of inputted hand-written Latin letter.

2. The identifying method of hand-written Latin letter, as recited in claim 1, wherein step (b) comprises the following steps of:

(b.1) smoothly filtering said chirography coordinates stored in the structural array, removing hardware noise, and eliminating burr and broken line formed in writing process;

(b.2) linearly and nonlinearly normalizing said chirography coordinate point lattice in the matrix format transformed from said chirography coordinates processed in step (b.1) by using linear normalization and non-linear normalization algorithm of on-line single-point width hand-written style to obtain new chirography coordinate point lattice; and

(b.3) adding points to said new chirography coordinate point lattice processed by step (b.2) to prevent from loosing effective characteristic point.

3. The identifying method of hand-written Latin letter, as recited in claim 2, wherein step (b.2) comprises steps of:

(b.2.1) linearly normalizing said chirography coordinate point lattice in the matrix format transformed from said chirography coordinates processed in step (b.1) to a standard coordinate point lattice by adopting coordinates frame scaling and direct projection method;

(b.2.2) expressing standard coordinate point of standard coordinate point lattice as (x,y), and classifying said standard coordinate point (x,y) into a chirography coordinate point and a non-chirography coordinate point;

(b.2.3) calculating a density function d(x,y) of said standard coordinate point (x,y),

$d<br><br>\left(x,y\right)=\{\begin{array}{c}\mathrm{Max}<br><br>\left(A/\mathrm{Lx},A/\mathrm{Ly}\right),\mathrm{Lx}+\mathrm{Ly}<br><br><br><br>\mathrm{\pi <br><br>}<br><br><br><br>6<br><br><br><br>A\\ 0,\mathrm{Lx}+\mathrm{Ly}\ge <br><br>6<br><br><br><br>A\end{array},$

wherein, Lx is a transverse density of a current standard coordinate point, Ly is a vertical density of a current standard coordinate point, A is a length of said standard coordinate point lattice, and a value of A is 64;

(b.2.4) calculating a level density projection function H(x) and a vertical density projection function V(y) of said standard coordinate point (x,y) according to said density function d(x,y) of said standard coordinate point (x, y),

$H<br><br>\left(x\right)=\underset{y=1}{\overset{J}{\sum <br><br>}}<br><br>\left[d<br><br>\left(x,y\right)+a_H\right],<br><br>V<br><br>\left(y\right)=\underset{x=1}{\overset{I}{\sum <br><br>}}<br><br>\left[d<br><br>\left(x,y\right)+a_v\right],$

wherein, d(x, y) is said density function of said standard coordinate point (x, y), I×

J is a dimension of said standard coordinate point lattice, a value of I×

J is 64×

64, a_H and a_v are correction coefficient, and both value of a_H and a_v are 0.1; and

(b.2.5) non-linearly normalizing said chirography coordinate point lattice by using a level density projection function H(x) and a vertical density projection function V(y) to obtain a new chirography coordinate point lattice,

$m=\underset{x=1}{\overset{i}{\sum <br><br>}}<br><br>H<br><br>\left(x\right)\times <br><br>\frac{M}{\underset{x=1}{\overset{I}{\sum <br><br>}}<br><br>H<br><br>\left(x\right)},<br><br>n=\underset{y=1}{\overset{j}{\sum <br><br>}}<br><br>V<br><br>\left(y\right)\times <br><br>\frac{N}{\underset{y=1}{\overset{J}{\sum <br><br>}}<br><br>V<br><br>\left(y\right)},$

wherein I×

J is the dimension of said standard coordinate point lattice, the value of I×

J is 64×

64;

M×

N is a dimension of said new chirography coordinate point lattice processed by non-linear normalization algorithm, a predetermined value of M×

N is 64×

64;

(m, n) is a coordinate point of said new chirography coordinate point lattice;

i=1,2,. . . , J, j=1,2, . . . ,J.

4. The identifying method of hand-written Latin letter, as recited in claim 1, wherein step (d) comprises steps of:

(d.1) determining whether the inputted hand-written Latin letter is up-down structure, wherein if the inputted hand-written Latin letter is up-down structure, determine whether the last stroke is transverse or not, wherein if the last stroke is transverse, directly determine that the coarse classification candidate letter aggregation matched with the inputted hand-written Latin letter is the transverse stroke coarse classification candidate letter aggregation;

if the last stroke is not transverse, directly determine that the coarse classification candidate letter aggregation matched with the inputted hand-written Latin letter is the non-transverse stroke coarse classification candidate letter aggregation;

if the inputted hand-written Latin letter is not up-down structure, carry out the following step;

(d.2) calculating the transverse cross number H, wherein when the transverse cross number is 1, calculate the vertical cross number, and determine the coarse classification candidate letter aggregation that matched with inputted hand-written Latin letter is the coarse classification candidate letter aggregation with S=1, S=2, S=3 or S=4 according to different vertical cross number;

when the transverse cross number is 2, calculate the transverse cross number UP in the ⅓

upper portion of the Latin letter, wherein when the transverse cross number in the ⅓

upper portion of the Latin letter is 2 or 3, directly determine that the coarse classification candidate letter aggregation matched with the inputted hand-written Latin letter is the coarse classification candidate letter aggregation with UP=2 or UP=3;

when the transverse cross number in the ⅓

upper portion of the Latin letter is 1, calculate the vertical cross number of the hand-written Latin letter, and determine the coarse classification candidate letter aggregation that matched with inputted hand-written Latin letter according to the vertical cross number and the position information of the last new chirography coordinate point of the new chirography coordinate lattice in the standard Latin letter during the hand writing, and whether the left arch opening of the inputted Latin letter existing or not;

when H=3, calculate the optimal vertical cross number of the inputted Latin letter, wherein when the optimal vertical cross number is 2 or 3, directly determine that the coarse classification candidate letter aggregation that matched with inputted hand-written Latin letter is the coarse classification candidate letter aggregation with Maxs=2 or Maxs=3;

when the optimal vertical cross number is 1, determine the coarse classification candidate letter aggregation that matched with inputted hand-written Latin letter according to the position information of the last new chirography coordinate point of the new chirography coordinate lattice in the standard Latin letter during the hand writing, and whether the left arch opening of the inputted Latin letter existing or not;

when H=4, calculate the transverse cross number Down in the ⅓

lower portion of the Latin letter, wherein when the transverse cross number in the ⅓

lower portion of the Latin letter is 2 or 3, directly determine that the coarse classification candidate letter aggregation matched with the inputted hand-written Latin letter is the coarse classification candidate letter aggregation with Down=2 or Down=3;

when the transverse cross number in the ⅓

lower portion of the Latin letter is 1, calculate the vertical cross number of the hand-written Latin letter, and determine the coarse classification candidate letter aggregation that matched with inputted hand-written Latin letter according to the vertical cross number and the position information of the last new chirography coordinate point of the new chirography coordinate lattice in the standard Latin letter during the hand writing, and whether the left arch opening of the inputted Latin letter existing or not.

5. The identifying method of hand-written Latin letter, as recited in claim 4, wherein step (d.1) comprises the following steps of (d.1.1) scanning the new chirography coordinate lattice row by row from top to bottom, wherein when meet with the first new chirography coordinate point, set the row value to the current row under scanning, and define the row value as the top boundary of the letter, and scanning the new chirography coordinate lattice row by row from bottom to top, wherein when meet with the first new chirography coordinate point, set the row value to the current row under scanning, and define the row value as the bottom boundary of the letter;

(d.1.2) scanning from the top boundary to the bottom boundary, record the number of the blank row where there is not new chirography coordinate point during the scanning, wherein when the number of the blank row equals to or is larger than 3 and next row of the last blank row of continuous blank rows is not blank, determine the inputted Latin latter is up-down structure.

6. The identifying method of hand-written Latin letter, as recited in claim 4, wherein in step (d.2), the calculation of the transverse cross number of the inputted hand-written Latin latter comprises the following steps of:

(d.2.1a) scanning each row of the new chirography coordinate lattice transversely, and recording the times of non new chirography coordinate point changing to new chirography coordinate point, wherein the changing times is the initial transverse cross number;

(d.2.2a) comparing the initial transverse cross number of each row in the new chirography coordinate lattice, and finding out the largest transverse cross number of the initial transverse cross numbers as the final transverse cross number of the letter;

in step (d.2), the calculation of the vertical cross number of the inputted hand-written Latin latter comprises the following steps of;

(d.2.1b) scanning each row of the new chirography coordinate lattice vertically, and recording the times of non new chirography coordinate point changing to new chirography coordinate point, wherein the changing times is the initial vertical cross number;

(d.2.2b) comparing the initial vertical cross number of each row in the new chirography coordinate lattice, finding out all the initial vertical cross numbers that equals to 3 or is larger than 3, and finding out the largest vertical cross number of the initial vertical cross numbers as the final vertical cross number of the letter;

in step (d.2), the calculation of the transverse cross number in the ⅓

upper portion and ⅓

lower portion of the hand-written Latin letter comprises the following steps of;

(d.2.1c) scanning the new chirography coordinate lattice row by row from top to bottom, wherein when meet with the first new chirography coordinate point, set the row value to the current row under scanning, and define the row value as the top boundary of the letter i1, and scan the new chirography coordinate lattice row by row from bottom to top, wherein when meet with the first new chirography coordinate point, set the row value to the current row under scanning, and define the row value as the bottom boundary of the letter i2;

(d.1.2c) calculating the ⅓

upper portion boundary i3, wherein i3=i1+⅓

(i2−

i1), and taking the rows between the top boundary i1 and the ⅓

upper portion boundary i3 as the ⅓

upper portion of the Latin letter;

calculating the ⅓

lower portion boundary i4, wherein i4=i1+⅔

(i2−

i), and take the rows between the bottom boundary i2 and the ⅓

lower portion boundary i4 as the ⅓

lower portion of the Latin letter;

(d.1.3c) according to the calculation process of the transverse cross number of whole Latin letter, obtaining the initial transverse cross number of each row from the top boundary i1 to the upper boundary row i3, finding out all the initial transverse cross numbers that equals to 3 or is larger than 3, and finding out the largest transverse cross number of the initial transverse cross numbers as the final transverse cross number in the ⅓

upper portion of the letter;

obtaining the transverse cross number in the ⅓

lower portion of the letter in the similar process mentioned above;

in step (d.2), the calculation of the optimal vertical cross number of the inputted hand-written Latin letter comprises the following step of;

(d.2.1d) during the calculation process of the vertical cross number of each row in the new chirography coordinate lattice, calculating the times of each initial vertical cross number, and taking the most frequently times as the optimal vertical cross number.

7. The identifying method of hand-written Latin letter, as recited in claim 4, wherein in step (d.2), the determining process of the position of the last new chirography coordinate point in the new chirography coordinate lattice comprises the following steps of:

(d.2.1e) scanning the new chirography coordinate lattice row by row from top to bottom, wherein when meet with the first new chirography coordinate point, set the row value to the current row under scanning, and define the row value as the top boundary of the letter i1, and scanning the new chirography coordinate lattice row by row from bottom to top, wherein when meet with the first new chirography coordinate point, set the row value to the current row under scanning, and define the row value as the bottom boundary of the letter i2;

scanning the new chirography coordinate lattice row by row from left to right, wherein when meet with the first new chirography coordinate point, set the row value to the current row under scanning, and define the row value as the left boundary of the letter j1, and scanning the new chirography coordinate lattice row by row from right to left, wherein when meet with the first new chirography coordinate point, set the row value to the current row under scanning, and define the row value as the right boundary of the letter j2, wherein the top boundary i1, bottom boundary i2, left boundary j1, and the right boundary j2 constitute the smallest outer rectangle frame of the hand-written Latin letter;

(d.2.2e) dividing the smallest outer rectangle frame evenly into a left-up rectangle frame, a left-down rectangle frame, a right-up rectangle frame, a right-down rectangle frame;

(d.2.3e) determining that the last new chirography coordinate point falls into the left-up rectangle frame, the left-down rectangle frame, the right-up rectangle frame or the right-down rectangle frame, wherein when the last new chirography coordinate point falls into the left-up rectangle frame, the last new chirography coordinate point is in the left-up position of the inputted hand-written Latin letter;

when the last new chirography coordinate point falls into the left-down rectangle frame, the last new chirography coordinate point is in the left-down position of the inputted hand-written Latin letter;

when the last new chirography coordinate point falls into the right-up rectangle frame, the last new chirography coordinate point is in the right-up position of the inputted hand-written Latin letter;

when the last new chirography coordinate point falls into the right-down rectangle frame, the last new chirography coordinate point is in the right-down position of the inputted hand-written Latin letter.

8. The identifying method of hand-written Latin letter, as recited in claim 6, wherein in step (d.2), the determining process of whether left arch opening existing in the hand-written Latin letter or not comprises the following steps of (d.2.1f) calculating the initial vertical cross number of each column in the new chirography coordinate lattice;

(d.2.2f) scanning from the right column to the left column, finding out the first column whose the initial vertical cross number is not 0, and recording the column as k;

(d.2.3f) when the initial vertical cross number of the (k-1)th column is larger than the initial vertical cross number of the kth column, defining the kth column as the reference column;

when the initial vertical cross number of the (k-1)th column is smaller than the initial vertical cross number of the kth column, stopping scanning and determine that there is no left arch opening in this letter;

when the initial vertical cross number of the (k-1)th column equals to the initial vertical cross number of the kth column, keeping scanning to the left until the initial vertical cross number of the (k-1)th column does not equal to the initial vertical cross number of the kth column, and defining the scanned column as the reference column;

(d.2.4f) when the reference column is determined, recording the vertical ordinate of the mth in-through point Im_in on the scanned reference column, that is the row ordinate of the in-through point Im_in in the new chirography coordinate lattice, recording the vertical ordinate of the mth out-through point Im_out, that is the row ordinate of the out-through point Im_out in the new chirography coordinate lattice, recording the vertical ordinate of the nth in-through point Jn_in on the scan line of the left column Ref_line_i-1 of next reference column, that is the row ordinate of the in-through point Jn_in in the new chirography coordinate lattice, and recording the vertical ordinate of the nth out-through point Jn_out on the scan line of the left column Ref_line_i-i of next reference column, that is the row ordinate of the out-through point Jn_out in the new chirography coordinate lattice, wherein m=1,2, . . . M, n=1,2, . . . N, M is the total number of the in-through point and out-through point of the scan line of the reference column, and N is the total number of the in-through point and out-through point of the scan line of the left column to the reference column;

(d.2.5f) finding out the middle point Im_mid between each couple of matched the in-through point Im_in and the out-thought point Im_out on the scan line of the reference column;

(d.2.6f) calculating the first straight distance between the each middle point Im_mid and the nth in-through point Jn_in, and finding out the in-through point on the scan line of the left column to the reference column that satisfy the first straight distance is less than or equals to 2;

calculating the second straight distance between the each middle point Im_mid and the nth in-through point Jn_out, and finding out the out-through point on the scan line of the left column to the reference column that satisfy the second straight distance is less than or equals to 2, wherein n=1,2, . . . N, and N is the total number of the in-through point and out-through point of the scan line of the left column to the reference column;

(d.2.7f) calculating the middle point between the pth in-through point on the scan line of the left column to the reference column that satisfy the first straight distance is less than or equals to 2 and the pth out-through point on the scan line of the left column to the reference column that satisfy the first straight distance is less than or equals to 2, and determining whether a point of same row with one of the calculated middle points on the scan line of the reference column is 0 or not, wherein when the point of same row with one of the calculated middle points on the scan line of the reference column is not 0, determine that there is left arch opening in the inputted handwritten Latin letter;

when the point of same row with one of the calculated middle points on the scan line of the reference column is 0, determine that there is no left arch opening in the inputted handwritten Latin letter;

wherein p=1, 2 . . . P, when the total number of the in-through points on the scan line on the left column to the reference column that satisfy the first straight distance is less than or equals to 2 is less than the total number of the out-through points on the scan line on the left column to the reference column that satisfy the second straight distance is less than or equals to 2, P is the total number of the in-through points on the scan line on the left column to the reference column that satisfy the first straight distance is less than or equals to 2;

when the total number of the in-through points on the scan line on the left column to the reference column that satisfy the first straight distance is less than or equals to 2 is more than the total number of the out-through points on the scan line on the left column to the reference column that satisfy the second straight distance is less than or equals to 2, wherein P is the total number of the out-through points on the scan line on the left column to the reference column that satisfy the second straight distance is less than or equals to 2.

9. The identifying method of hand-written Latin letter, as recited in claim 1, wherein step (e) comprises the following steps of:

(e.1) extracting the freeman chain code of the inputted hand-written Latin letter;

(e.2) calculating the matching probability between the freeman chain code of the inputted hand-written Latin letter and the optimal Hidden Markov Models of each standard letter stored in a predetermined letter library and corresponding to each letter stored in the coarse classification candidate letter aggregation by using Viterbi Algorithm;

(e.3) obtaining the maximum matching probability by sequencing each matching probability by using public Bubble Sort Algorithm;

(e.4) determining the standard letter stored in the predetermined letter library and corresponding to the optimized matching probability as the final identifying result of Latin letter.