CN100585623C - Method and system for discriminating handwriting characters - Google Patents
Method and system for discriminating handwriting characters Download PDFInfo
- Publication number
- CN100585623C CN100585623C CN200710017598A CN200710017598A CN100585623C CN 100585623 C CN100585623 C CN 100585623C CN 200710017598 A CN200710017598 A CN 200710017598A CN 200710017598 A CN200710017598 A CN 200710017598A CN 100585623 C CN100585623 C CN 100585623C
- Authority
- CN
- China
- Prior art keywords
- stroke
- handwriting characters
- handwriting
- section
- characters
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Landscapes
- Character Discrimination (AREA)
Abstract
A method for identifying character of hand-writing input carry out treatment on inputted stroke by utilizing corner cut of included angle between each adjacent stroke sections in character of hand-writing input for adjusting position and structure between stroke sections so as to pick up feature information of character inputted by hand-writing for carrying out identification.
Description
Technical field
The present invention relates to the character recognition system and method, particularly a kind of discriminating handwriting characters method and system, this method and system thereof can according in the handwriting characters between each adjacent strokes section the corner cut of angle to its input stroke be optimized processing, to adjust intersegmental structure of stroke and position, realize discriminating handwriting characters.
Background technology
Along with developing rapidly of science and technology, various palm-type data processors constantly occur, as: mobile phone, personal digital assistant hand-held electronic products such as (PDA) nearly all possess the Handwritten Digits Recognition function.Usually the literal operating characteristic is captured by input equipment, and operating characteristic can be any useful structural features information or its combination in the sampling data, and wherein basic structural feature comprises cellular construction features such as unique point, stroke section and stroke.
But, because input equipment or human factor can make the character of input comprise many error noises that cause that identification makes mistakes.Common disposal route is, falls by the filter algorithm error noise filtering of stroke in handling, and only keeps the unique point and the stroke segment information of stroke.Described filter method is only determined point that the direction value is close for fixing stroke section with the direction tendency of stroke mid point, if the stroke tortuosity is bigger, then will be divided into a lot of short stroke sections when the bigger stroke of radian of input.Originally can be filtered fully, and can keep too much stroke section, in template information, also show as the feature data and become very big by this filter method in same direction stroke section.In addition, keeping too much unique point can increase error because the last stroke Duan Yuhou one stroke section of comparison needs reference point, therefore make aspect ratio to the time increase unnecessary reference point influence.Therefore, the present most algorithm only stroke section for more straight is effective, such as Chinese character etc., and for such as excessive stroke DeGrains of tortuosity such as safe literal unit and multiple lexigraphies.Because the stroke section and the unique point that are kept after filtration treatment are too much, and then eigenwert is increased, when the comparison template, cause identification speed slow, cause discrimination power low excessively.
Summary of the invention
Problem and defective at above-mentioned prior art existence, the object of the invention is, a kind of discriminating handwriting characters method and system are provided, and the influence that input equipment identifies the handwriting and the error of human factor can be removed by this method and system thereof, improve the processing speed and the discrimination power of handwriting identification.
In order to realize above-mentioned task, the present invention takes following technical solution:
A kind of discrimination method of handwriting characters is characterized in that, this method is carried out identification to handwriting characters at first then, comprises the following steps:
Obtain a hand-written input character of being formed by a plurality of stroke sections, the stroke of handwriting characters is carried out error noise filtering, filtering, remove the noise that should not occur in the character picture;
Calculate the corner cut of angle between each adjacent strokes section of handwriting characters then, by the corner cut of the angle that forms between each adjacent strokes section of handwriting characters the input stroke is handled, to adjust intersegmental structure and the position of stroke;
Judge that continuously described corner cut draws the position to find out a long straight pen, determine the slope of the Origin And Destination connection that long straight pen is drawn, wherein long straight pen draw on Duan Weiyi the direction and comprise stroke section more than three at least;
Deletion is drawn the stroke section of slope difference minimum with long straight pen, and connects remaining stroke section in the handwriting characters; And
And then the characteristic information of acquisition handwriting characters calculates, provide stroke section characteristic sequence, and browse all template informations in the template database, and make result of calculation and a standard form wherein compare, obtain the final identification result of handwriting characters.
Realize the discriminating handwriting characters system of said method, this system comprises:
One input equipment is used to receive a hand-written input character;
One pretreatment module is used for handwriting characters is carried out error noise filtering, filtering, removes the noise that should not occur in the character picture;
One template database is used to store the plurality of standard template;
One identification module, being used for that handwriting characters stroke section is carried out characteristic information calculates, and a standard form of characteristic information and template database compared, to adjust intersegmental structure and the position of stroke, the characteristic information of acquisition handwriting characters carries out discriminating handwriting characters, and identification result is saved to storage device or shows identification result by display device;
Input equipment is connected with identification module by pretreatment module, and template database is connected on the identification module, and identification module also links to each other with display device with storage device respectively.
The discrimination method of input character of the present invention and system, use the corner cut of stroke each several part stroke section angle to calculate and handle intersegmental structure of stroke and position, unreasonable factor in the known discrimination method is carried out optimization to be handled, can remove the influence that input equipment identifies the handwriting and the error of human factor, reduce the generation of System Discrimination error.In addition, as filtering, make characteristic reduce, remove more error point effectively, and then better comparison mechanism is provided, reduce the identification error, improved the processing speed and the discrimination power of handwriting identification according to filtering out more redundant stroke section with corner cut.
Description of drawings
Fig. 1 is a discriminating handwriting characters system block diagram of the present invention;
Figure is 2 discriminating handwriting characters method flow diagrams of the present invention;
Fig. 3 is the stroke process method flow diagram of handwriting characters of the present invention;
Fig. 4 is for being a Processing Example of example with safe literal body, and wherein figure (a) is the safe literal body of a regular body synoptic diagram; Figure (b) is the hand-written safe literal body person's handwriting synoptic diagram of user; Figure (c) is for handling the resulting identification output in back display graphics by conventional filter algorithm to the hand-written safe literal body person's handwriting shown in the figure (a), stroke process method the identification output display graphics to figure (b) shown in hand-written safe literal body person's handwriting handle after obtain of figure (d) for adopting handwriting characters of the present invention.
Label among the above-mentioned figure is represented respectively: 10, discriminating handwriting characters system, 12, input equipment, 14, pretreatment module, 16, identification module, 18, template database, 20, display device, 22, storage device;
The present invention is described in further detail below in conjunction with embodiment that accompanying drawing and inventor provide.
Embodiment
Principle of the present invention is to utilize the corner cut of adjacent two stroke section institute angles, as filtering according to removing unnecessary error character data, to improve discrimination power and speed.
Referring to Fig. 1, this figure is the calcspar of a kind of discriminating handwriting characters system that provides of the present invention, as shown in the figure, this discriminating handwriting characters system 10 comprises: an input equipment 12, a pretreatment module 14, an identification module 16 and a template database 18, the user obtains handwriting characters by writing words on the handwriting pad of input equipment 12, then input equipment 12 all the handwriting characters intelligence samples on the handwriting pad after send in the pretreatment module 14 by the serial or parallel interface.14 pairs of handwriting characters of pretreatment module comprise the pretreatment operation of error noise filtering, filtering, and then with the noise remove that should not occur in the character picture.Then 16 pairs of pretreated handwriting characters of identification module carry out the characteristic information acquisition, here, recognition module 16 is when the characteristic information of acquisition handwriting characters, at first obtain the handwriting characters of being formed by a plurality of stroke sections, by in the handwriting characters between each adjacent strokes section the corner cut of angle the stroke of handwriting characters is handled, to adjust intersegmental structure and the position of stroke, promptly, calculate the corner cut of angle between each adjacent described stroke section of handwriting characters, the judgement corner cut is drawn the place to find out a long straight pen continuously, long straight pen is drawn by on the direction and stroke section that comprise at least more than three and is formed, determine the slope of the Origin And Destination connection that long straight pen is drawn then, deletion and long straight pen are drawn the stroke section of slope difference minimum, remaining stroke section are connected the feature stroke segment information that promptly obtains handwriting characters.Then, 16 pairs of handwriting characters characteristic informations of recognition module calculate, and provide the characteristic sequence of stroke section, and browse all template informations in the template database 18, make result of calculation and a standard form wherein compare, and obtain final identification result.At last identification result is saved in the storage device 22, perhaps shows the identification result of handwriting characters by a display device 20.By aforesaid operations the handwriting characters stroke is handled, with the corner cut of the angle of adjacent two stroke sections as filtering foundation, can on vertical and horizontal direction, make the feature of available stroke section more can embody the characteristic of literal, filter the error that writing style or input equipment because of different people cause, and then having improved the comparison degree of accuracy, the minimizing of characteristic value information amount has also improved the speed when comparing.
Referring to Fig. 2, this figure is the discriminating handwriting characters method flow diagram of Fig. 1, at first handwriting characters is carried out pretreatment operation, promptly input character is carried out error noise filtering, Filtering Processing (step 101), by the corner cut of the angle that forms between each adjacent strokes section of handwriting characters the input stroke is handled then, with intersegmental structure and position (step 102) of adjustment stroke, and then the characteristic information of acquisition handwriting characters carries out discriminating handwriting characters (step 103).When carrying out discriminating handwriting characters, need to calculate the stroke section feature of handwriting characters, and the characteristic sequence that provides the stroke section is described (step 104) to be identified, in step 105, browsing template information compares to characteristic information, wherein need to calculate the distance size of the characteristic information and the image information features in each template information of handwriting characters, the more little expression comparison rate of distance is high more, then according to distance ordering output identification result (step 106).
Below in conjunction with Fig. 3 and Fig. 4 a to Fig. 4 d the stroke process method of handwriting characters of the present invention is described.
Fig. 3 is the stroke process method flow diagram of handwriting characters of the present invention, i.e. the concrete steps of the realization of the step 102 described in Fig. 2.Fig. 4 a to Fig. 4 d is the detailed maps of discriminating handwriting characters method one embodiment of the present invention, wherein Fig. 4 a is the safe literal body of a regular body synoptic diagram, Fig. 4 b is the hand-written safe literal body person's handwriting synoptic diagram of user, Fig. 4 c is for handling the resulting identification output in back display graphics by known filter algorithm to hand-written safe literal body person's handwriting shown in Fig. 4 b, wherein 1., 2., 3. the pen section of locating causes interference for error pen section to identification, in the filtering since this section proportion in character stroke be retained more greatly down, this error can increase data after the feature extraction, reduce the hit rate of identification template, identification speed is slowed down.And stroke process method the identification output display graphics to Fig. 4 b shown in hand-written safe literal body person's handwriting handle after obtain of Fig. 4 d for adopting handwriting characters of the present invention.When utilizing known filter algorithm to carry out discriminating handwriting characters, generally speaking feature extraction is to carry out unique point, stroke section or/and the stroke information extraction after the pre-service of handwriting characters stroke, but above-mentioned feature extraction is for the excessive stroke DeGrain of tortuosity.As shown in Figure 3, discriminating handwriting characters method of the present invention by in the handwriting characters between each adjacent strokes section the corner cut of angle to its input stroke handle, to adjust intersegmental structure and the position of stroke, at first obtain the handwriting characters of being formed by a plurality of stroke sections (step 201), pass through the slope (step 202) between the coordinate Calculation two stroke sections of each adjacent strokes section then, and then obtain the coordinate quadrature (step 203) at two stroke section places.Calculate the complementary angle of two stroke section angles by the quadrature and the slope meter of two stroke sections then, and then obtain its corner cut (step 204).After the corner cut that calculates all adjacent two stroke section angles, judge each continuous corner cut, maximum angular is a deflection angle, and then finds out the place that long straight pen is drawn, be on the direction and comprise stroke section (step 205) more than three at least.Obtain long straight pen and draw middle starting point and online slope of terminal point and the quadrature (step 206) in coordinate system thereof, deletion and long straight pen are drawn the stroke section of slope difference minimum at last, and remaining stroke section promptly obtains the characteristic information (step 206) of the handwriting characters shown in Fig. 4 d in the connection handwriting characters.From the shown figure of Fig. 4 c and Fig. 4 d as can be seen, the stroke process method of handwriting characters of the present invention not only can filter out more redundant stroke section, the characteristic information of handwriting characters be can highlight again, and then the processing speed and the discrimination power of identification improved.
Though the present invention has made detailed description with above-mentioned preferred embodiment, be not to limit the present invention with the foregoing description.Those skilled in the art should recognize under the situation that does not break away from given technical characterictic of technical solution of the present invention and scope, and the increase that technical characterictic is done or with the replacement of some same contents of this area all should belong to protection scope of the present invention.
Claims (2)
1. the discrimination method of a handwriting characters is characterized in that, this method is at first carried out identification to handwriting characters, comprises the following steps:
Obtain a hand-written input character of being formed by a plurality of stroke sections, the stroke of handwriting characters is carried out error noise filtering, filtering, remove the noise that should not occur in the character picture;
Calculate the corner cut of angle between each adjacent strokes section of handwriting characters then, by the corner cut of the angle that forms between each adjacent strokes section of handwriting characters the input stroke is handled, to adjust intersegmental structure and the position of stroke;
Judge that continuously described corner cut draws the position to find out a long straight pen, determine the slope of the Origin And Destination connection that long straight pen is drawn, wherein long straight pen draw on Duan Weiyi the direction and comprise stroke section more than three at least;
Deletion is drawn the stroke section of slope difference minimum with long straight pen, and connects remaining stroke section in the handwriting characters; And
And then the characteristic information of acquisition handwriting characters calculates, provide stroke section characteristic sequence, and browse all template informations in the template database, and make result of calculation and a standard form wherein compare, obtain the final identification result of handwriting characters.
2. enforcement of rights requires the system of 1 described discriminating handwriting characters method, it is characterized in that, this system comprises:
One input equipment is used to receive a hand-written input character;
One pretreatment module is used for handwriting characters is carried out error noise filtering, filtering, removes the noise that should not occur in the character picture;
One template database is used to store the plurality of standard template;
One identification module, obtain the handwriting characters of being formed by a plurality of stroke sections, by in the handwriting characters between each adjacent strokes section the corner cut of angle the stroke of handwriting characters is handled, adjust intersegmental structure and the position of stroke, calculate the corner cut of angle between each adjacent described stroke section of handwriting characters, the judgement corner cut is drawn the place to find out a long straight pen continuously, long straight pen is drawn by on the direction and stroke section that comprise at least more than three and is formed, determine the slope of the Origin And Destination connection that long straight pen is drawn, deletion and long straight pen are drawn the stroke section of slope difference minimum, remaining stroke section is connected the feature stroke segment information that promptly obtains handwriting characters, handwriting characters stroke section is carried out characteristic information to be calculated, and a standard form of characteristic information and template database compared, the characteristic information of acquisition handwriting characters carries out discriminating handwriting characters, and identification result is saved to storage device or shows identification result by display device;
Input equipment is connected with identification module by pretreatment module, and template database is connected on the recognition template, and recognition template also links to each other with display device with storage device respectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200710017598A CN100585623C (en) | 2007-03-30 | 2007-03-30 | Method and system for discriminating handwriting characters |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200710017598A CN100585623C (en) | 2007-03-30 | 2007-03-30 | Method and system for discriminating handwriting characters |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101030261A CN101030261A (en) | 2007-09-05 |
| CN100585623C true CN100585623C (en) | 2010-01-27 |
Family
ID=38715588
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200710017598A Expired - Fee Related CN100585623C (en) | 2007-03-30 | 2007-03-30 | Method and system for discriminating handwriting characters |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100585623C (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101833411B (en) * | 2009-03-09 | 2015-09-16 | 诺基亚公司 | For the method and apparatus of person's handwriting input |
| CN102169413A (en) * | 2011-03-30 | 2011-08-31 | 黄冬明 | Device and method for obtaining character stroke lines based on video stream image |
| JP6229303B2 (en) * | 2013-05-16 | 2017-11-15 | 富士通株式会社 | Program, information processing apparatus, and character recognition method |
| CN103778250B (en) * | 2014-02-19 | 2017-02-15 | 张朝亮 | Implement method for universal Chinese wubi cursive script dictionary query system |
| CN106776656A (en) * | 2015-11-25 | 2017-05-31 | 中兴通讯股份有限公司 | Adjust the display methods and device of picture character |
-
2007
- 2007-03-30 CN CN200710017598A patent/CN100585623C/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN101030261A (en) | 2007-09-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| RU2711029C2 (en) | Touch classification | |
| US10013624B2 (en) | Text entity recognition | |
| US8315482B2 (en) | Integrated platform for user input of digital ink | |
| US8180160B2 (en) | Method for character recognition | |
| US20200065601A1 (en) | Method and system for transforming handwritten text to digital ink | |
| CN109685052A (en) | Method for processing text images, device, electronic equipment and computer-readable medium | |
| CN100585623C (en) | Method and system for discriminating handwriting characters | |
| CN102184395B (en) | String-kernel-based hand-drawn sketch recognition method | |
| US9378427B2 (en) | Displaying handwritten strokes on a device according to a determined stroke direction matching the present direction of inclination of the device | |
| WO2013184726A2 (en) | Adaptive thresholding for image recognition | |
| JP2007317022A (en) | Handwritten character processor and method for processing handwritten character | |
| WO2015090075A1 (en) | Touch screen terminal and method and system for identifying hand gestures of touch screen terminal | |
| US8363947B2 (en) | Handwritten character recognition based on frequency variations in characters | |
| CN102982155A (en) | System and method for searching defects based on pattern identification | |
| CN110147785B (en) | Image recognition method, related device and equipment | |
| CN103235836A (en) | Method for inputting information through mobile phone | |
| CN105426877A (en) | Method and system for information recognition applying sweat gland position information | |
| JP5735126B2 (en) | System and handwriting search method | |
| CN109101973A (en) | Character recognition method, electronic equipment, storage medium | |
| CN112183149B (en) | Graphic code processing method and device | |
| Nguyen et al. | Semi-incremental recognition of on-line handwritten Japanese text | |
| CN100394435C (en) | Word recognition method and its system | |
| Amarnath et al. | Automatic localization and extraction of tables from handheld mobile-camera captured handwritten document images | |
| CN116414289B (en) | Gesture recognition system and method based on touch chip | |
| WO2014181433A1 (en) | Electronic device, handwritten document search method, and program |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20100127 Termination date: 20110330 |