CN104898855B - Based on text input system and method with rocking bar equipment - Google Patents
Based on text input system and method with rocking bar equipment Download PDFInfo
- Publication number
- CN104898855B CN104898855B CN201510305823.XA CN201510305823A CN104898855B CN 104898855 B CN104898855 B CN 104898855B CN 201510305823 A CN201510305823 A CN 201510305823A CN 104898855 B CN104898855 B CN 104898855B
- Authority
- CN
- China
- Prior art keywords
- input
- handwriting
- svg
- word
- person
- 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
Landscapes
- Character Discrimination (AREA)
Abstract
It is a kind of based on text input system and method with rocking bar equipment, including：Input equipment, handwriting characteristic extraction module, person's handwriting SVG models, handwriting model training module, dictionary association algoritic module, interface information control module, display device, wherein：Input equipment be connected with handwriting characteristic extraction module and transmit user input person's handwriting a series of plane coordinates, handwriting characteristic extraction module and handwriting model training module are connected with person's handwriting SVG models and input person's handwriting characteristic information and input results feedback information respectively, person's handwriting SVG models associate algoritic module and interface information control module with dictionary and are connected and export the result that person's handwriting identifies respectively, dictionary association algoritic module is connected and exported under association's pattern the most possible word filtered out with interface information control module, interface information control module is connected with display device and exports animation information, feedback information is passed to handwriting model training module by user according to input results correctness, for training new person's handwriting SVG models, so as to improve constantly the accuracy rate of identification, the present invention can significantly improve input efficiency.
Description
Technical field
It is specifically a kind of based on rocking bar equipment the present invention relates to a kind of technology of field of computer peripherals
Text input system and method.
Background technology
It is sitting on sofa when using Xbox or intelligent television, inputs the word of several similar user names or movie title conveniently
Mother is very common thing.The demand for carrying out word input using rocking bar on game host and intelligent television exists always, and with
Greatly enriching for internet content and quickly increase.The new technologies such as speech recognition are as a kind of optional word input scheme
Traditional inputting interface can not be still replaced under many circumstances, such as keyboard and rocking bar.Increase on intelligent television or game host
The words such as long registration, search, short message, Email input demand, a kind of effective word input medium will substantially improve this
The experience of a little functions, the also further development for short message and Email are laid a good foundation.
Rocking bar kind equipment is widely used as a kind of entity control in intelligent television, game and vehicle entertainment system, its
Contain rocking bar or its variant form, including but not limited to game paddle rocking bar, handheld device rocking bar, the rotation of vehicle-mounted Comprehensive Control
Button, the multi-direction button of annular.
The method that word input is carried out using rocking bar main at present is cursor selection virtual keyboard input method.This method is
One piece of dummy keyboard is shown on screen, is inputted using the letter on rocking bar movement cursor selection keyboard, wherein：, letter
Table keyboard and Qwerty keyboards are most popular keyboard arrangement modes.The character input method is widely used in game machine (such as rope
The PSP handheld and PS main frames, the Xbox main frames of Microsoft etc. of Buddhist nun company), intelligent television, vehicle-mounted information and entertainment system (such as BMW
IDrive, benz COMAND, Audi MMI etc.).The advantages of this input mode be study threshold it is low, shortcoming be it is less efficient, it is defeated
Enter that process is dry as dust, also, due to needing user to keep the notice to dummy keyboard, this input method can not possibly be realized blind
Input.
Foreign countries have researcher to propose similar to hand-written new input way, for example, Graffiti, Unistrokes and
EdgeWrite.Rocking bar is difficult to the complex plane track of accurate " writing " letter, therefore these input methods pair due to its physical property
Letter simplify to facilitate writing.Graffiti and Unistrokes is the use stylus proposed the 1990s
Carry out hand-written input method.They are easy to write, improve the effect of machine recognition by the way that letter is simplified into a stroke.
EdgeWrite adds the physical edge auxiliary rocking bar writing of a square frame.The font of each letter contains a system
Angle and side on row square frame.Compared with tradition selects inputting method, these input methods relatively select inputting method to have can be blind defeated
The advantage entered.But user needs to learn new alphabet, therefore input speed at the beginning is slow.CHI meetings in 2007
On Game Controller Text entry with Alphabetic and Multi-Tap Selection
Keyboard articles propose the input for controlling two keyboards to be inputted respectively using two rocking bars on Xbox360 handles
Method.The input method is each responsible for two different zones based on user using right-hand man during input through keyboard, and qwerty keyboard is divided into
Two pieces improve input efficiency.But still need user's holding to the notice of dummy keyboard, blind input can not be realized.
Found by the retrieval to prior art, Chinese patent literature CN1607491 discloses (bulletin) day
2005.04.20, a kind of Chinese version word input system and method are disclosed, it is allowed to which user is by using control stick or its is same
Deng thing, word is input in the device of such as mobile phone or PDA by the first few stroke of addition word needs.Due to simply
The addition of mobile operating bar is used for starting one or more strokes of written word, or sometimes before any stroke is added,
User can also find a desired word from the selection table of display.The selection table is context-sensitive, by last time
The word of input and it is different, so as to user can have maximum possible candidate desired word.But the prior art and this hair
Bright to compare, its insurmountable technical problem includes 31 kinds of stroke forms of Chinese character and 5 kinds of control stick operate morphologically not
It can correspond to completely, novice users need to remember, and study threshold is higher；The tilted direction movement of control stick is unmanageable, in this patent
In be moved to 4：30 directions and 7：Easy maloperation during 30 direction；Lack the mechanism for helping user to learn in use；During operation
Feedback can not be seen on software interface, can not be in software interface when such as carrying out a stroke input using control stick
To the stroke specifically inputted.
Chinese patent literature CN1567160 discloses (bulletin) day 2005.01.19, discloses a kind of input device
And its operation method, input device include：One direction device, to detect and export a plurality of directions, it is intended to input a word
When according to one of those directions corresponding to the output of character stroke handle bar device for steering；One control circuit is coupled to direction device, will
One of those directions coding of direction device output and storage, and be compared with database, and export those corresponding directions it
One plural candidate word.The operation method includes：Receive one direction of input；By direction encoding and storage；Enter with a database
Row compares, and exports a plurality of candidate characters of the corresponding direction.The technology utilizes Joystick-type or the direction device of steering-wheel type
Input word, it is only necessary to a finger control direction can utilization orientation information by word according to stroke depict and reach input text
Word result；Need not use input through keyboard that direction device is used only coordinates Input Software to reach word input and can save sky
Between；Only with refer on the other hand control direction and input word need not departure direction device, therefore use simple.But the prior art and this hair
Bright to compare, its insurmountable technical problem includes well not corresponding between character stroke and direction, it is difficult to remembers；Some sides
It is difficult to accurately export 8 directions, particularly tilted direction to output device such as rocking bar；Lack picture feedback；Lack and help user to learn
Practise the mode used.
The content of the invention
The present invention is directed to deficiencies of the prior art, proposes a kind of based on the text input system with rocking bar equipment
And method, it can conveniently realize blind defeated, make use of the learning difficulty of rocking bar low so that the present invention can significantly improve input effect
Rate.
The present invention is achieved by the following technical solutions：
The present invention relates to a kind of based on the text input system with rocking bar equipment, including：Input equipment, handwriting characteristic extraction
Module, person's handwriting SVG models, handwriting model training module, dictionary association algoritic module, interface information control module, display device,
Wherein：
Input equipment be connected with handwriting characteristic extraction module and transmit user input person's handwriting a series of plane coordinates, pen
Mark characteristic extracting module and handwriting model training module are connected with person's handwriting SVG models and input person's handwriting characteristic information and defeated respectively
Enter result feedback information, person's handwriting SVG models associate algoritic module with dictionary respectively and interface information control module is connected and exported
The result of person's handwriting identification, dictionary association algoritic module, which is connected with interface information control module and exported under association's pattern, to be filtered out
Most possible word, interface information control module is connected with display device and exports animation information, and user ties according to input
Feedback information is passed to handwriting model training module by fruit correctness, for training new person's handwriting SVG models, so as to improve constantly
The accuracy rate of identification.
Under association's input pattern, person's handwriting SVG models and dictionary association algoritic module are connected and transmit the knot that person's handwriting identifies
Fruit, it is possible to letter and its probability；Under non-association's pattern, person's handwriting SVG models are directly connected with interface information control module,
And the result for transmitting person's handwriting identification selects for user, it is no more than 3 according to the quantity of the optional result of probability size.
Described person's handwriting SVG models, the handwriting trace of user and correct recognition result are entered by on-line study mechanism
Row learning classification, to be optimized to SVG models, so as to further improve the accuracy rate of character classification.
The present invention relates to the text entry method of said system, comprise the following steps：In being moved first by using rocking bar
Rocking bar touching or leave border, rocking bar returns to center and reverse three kinds of states split track and according to curvature threshold on border
Value segmentation characteristic vector pickup track characteristic, and the SVG models for the handwriting trace of Real time identification user are trained to, then use
SVG models carry out character classification judgement to the same rocking bar input trajectory for carrying out feature extraction, and export a series of possible words
Female and its probability.
Methods described is preferably cut between association's input pattern (whole word input) and non-association's input pattern (letter-by-letter input)
Change, under association's input pattern, the character classification judged result of all tracks inputted will calculate with the word in dictionary
Joint probability provides probability highest word, and is constantly updated with input is continued.Under non-association's input pattern, probability is higher than
The character judged result of threshold value and quantity no more than three will be supplied to user to be selected.
Described method, specifically includes following steps：
Step 1, collection initial user letter track sample characteristics simultaneously train SVG learning models, specifically include：
1.1 read the plane coordinates of alphabetical handwriting trace；
1.2 moved according to rocking bar in rocking bar touching or leave border, rocking bar returns to center and reverse three kinds on border
Whether condition adjudgement adds new stroke, and each stroke is made up of a series of coordinate points；
1.3 screen according to sampled point quantity to stroke, remove noise；
Each stroke in 1.4 tracks be split further according to certain curvature threshold for be similar to the feature of straightway to
Amount；
1.5 from the multinomial feature of each characteristic vector pickup, including distance, angle, absolute position, absolute angle, definitely away from
From and skew；
1.6 train SVG models according to track characteristic set.
Step 2, a rocking bar input trajectory is identified, specifically included：
A 2.1 rocking bar input trajectory coordinate datas of collection；
2.2 extraction track characteristics；
2.3 import SVG learning models；
2.4 output recognition results, comprising multiple letters and its probability, when present mode is associates input pattern, enter
Step 3, otherwise into step 4；
Step 3, according to track identification result dictionary is screened, specifically included：
3.1 remove the word that possibility is lost in dictionary according to this recognition result；
3.2 judge now whether dictionary is empty；
If 3.3 dictionaries, which are sky, provides prompting, if not being superimposed impossible word in alphabetical group for empty remove in input field
It is female.
Letter is inputted in step 4, selection candidate list, is specifically included：
4.1 screen according to probability threshold values to recognition result, only retain the letter that probability is more than threshold values；
4.2 show probability highest letter by default, and in addition first three recognition result of probability is as candidate list；
If there is no target alphabetical in 4.3 acquiescences or candidate list, delete current results and re-enter, otherwise continue next
Step；
The selected letter of 4.4 users；
The selected letter of 4.5 input fields display input；
4.6 are successfully entered after letter according to alphabetical and corresponding input trajectory renewal SVG models.
Step 5, selection word continue input letter, specifically include：
5.1 calculate joint using HMM algorithms according to the possible result of each letter and the word frequency of probability and word in word
Probability；
5.2 are arranged word in dictionary as candidate list by probability from big to small, and picture default word and word are waited
Table is selected to be updated；
If 5.4 select word from candidate list, the word is inputted；If do not selected, continue to input alphabetical track,
Repeat step 1~4, until selected word；
5.4 words input successfully after according to alphabetical and corresponding input trajectory update SVG models.
Technique effect
Compared with prior art, the present invention improves the efficiency that English input is carried out using rocking bar kind equipment.On the one hand, it is right
Very low in the study threshold of novice users, input efficiency is higher at the very start.On the other hand, on-line study mechanism is continuous more
The writing sample of new user so that for a long time in use, system more conforms to user writing custom, input speed and accuracy rate are big
Width is lifted.
Brief description of the drawings
Fig. 1 is the exemplary function structure chart of input method system；
Fig. 2 is the algorithm flow schematic diagram of extraction characteristic module；
Fig. 3 is the algorithm flow schematic diagram of training SVG models；
Fig. 4 is the algorithm flow schematic diagram of association's input pattern；
Fig. 5 is the algorithm flow schematic diagram of non-association's input pattern；
Fig. 6 is a kind of schematic diagram of the possible rocking bar equipment Xbox360 handles using the input system；
Fig. 7 is a kind of schematic diagram of the possible rocking bar equipment PSP game machines using the input system；
Fig. 8 is a kind of schematic diagram of possible rocking bar equipment TV/set-top box remote controller using the input system；
Fig. 9 is a kind of schematic diagram of the possible rocking bar equipment vehicle mounted guidance knob using the input system；
Figure 10 is to write letter using rocking bar and the schematic diagram of the track of letter is write using touch pad；
Figure 11 is the schematic diagram of the English inputting interface of the present invention.
Embodiment
Embodiments of the invention are elaborated below, the present embodiment is carried out lower premised on technical solution of the present invention
Implement, give detailed embodiment and specific operating process, but protection scope of the present invention is not limited to following implementation
Example.
Embodiment 1
A kind of as shown in fig. 6, rocking bar kind equipment Xbox360 handles that this implementation system is supported.Have two on Xbox360 handles
Individual rocking bar, user can be write using one of rocking bar, be selected using another rocking bar.In user test,
The rocking bar in the upper left corner has been used to be used to write, the rocking bar in the lower right corner is selected, and can be actually accustomed to setting according to user.Handle
On three buttons X, A, B be each set to delete, determine and associate mode switch function.Figure 11 example inputting interfaces include
Input field and association's candidate bar, input field are used for the letter after showing feedback animation and determining, association's candidate bar is in association's pattern
In be used for select possible word, be not present in non-association's pattern.
The identification method write using rocking bar has used for reference touch pad to a certain extent.But both still have it is huge
Big difference.As shown in Figure 10, the track on touch pad spreads out in one plane and has obvious breakpoint, and uses rocking bar
The track of writing is typically continuous, and many strokes can be overlapped on border.Because when being write using rocking bar, abut
Physical boundary movement is most natural and full blast.In order to split stroke, after track is divided into several strokes, then it is right respectively
Each stroke part carries out cutting according to curvature threshold values.Therefore each track can be subdivided into multiple strokes, and each stroke is again
It is made up of a series of near linear section.For each near linear section, 7 features, including distance, angle are all therefrom extracted
Degree, absolute position, absolute angle, absolute distance and skew.Certainly because a letter may have a variety of literary styles, therefore each word
Mother has multiple corresponding tracks.
The feature extracted from the writing sample of user is used to be trained to a SVG (Scalable Vector
Graphics) model, the model is by the handwriting trace for Real time identification user.For sample supplier, this system
It is being more accurately at the very start.For other users, it is found that the degree of accuracy of the diversity of sample to system has
Significant impact.Except allowing sample to cover more possible writing styles as far as possible, the on-line study mechanism of proposition can be with user
Input gradually step up the recognition effect of system.Two kinds of input patterns are provided in system：Associate input pattern and non-association is defeated
Enter pattern.
In association's input pattern, user is inputted by word.For each alphabetical track of user writing, SVG models
Multiple possible recognition results and corresponding probability will be provided.Meeting of the accuracy rate less than threshold values is arranged in the letter identified
Remove, the higher letter of remaining accuracy rate is presented at input field cursor after the screening of HMM algorithms in a manner of superposition.One
In the input process of individual word, user often inputs a letter, HMM algorithms can all be inputted according to user in whole word one
Word present in the recognition result and dictionary of serial track calculates joint probability, and the word result that probability is 0 can be excluded,
It is exactly that letter disappears accordingly in alphabetical group be superimposed to be reflected on picture, and remaining possible word is arranged from big to small by probability
In association's column candidate bar.Therefore the whole input process of a word is exactly that the result for just having started to input every time is all superimposed upon
Together alphabetical group, and as input alphabetical one by one, the letter of superposition gradually decrease, the target word of user gradually emerges
Out.Target word emerge speed and word property in itself and the size of dictionary is related.When a word input successfully,
Line study mechanism will learn each alphabetical and corresponding track in word, SVG models is more conformed to the writing of user
Custom, improve recognition accuracy.
In non-association's pattern, user often inputs a track, and 3 letters of probability highest can be by the result identified
Probability size order is arranged in that input cursor is inferior to treat that user selects, and after user selects some letter, corresponding letter is just completed
Input.The alphabetical and corresponding track will be brought into SVG models by on-line study mechanism, improve the knowledge for the user
Other accuracy rate.
Claims (6)
- It is 1. a kind of based on the text input system with rocking bar equipment, it is characterised in that including：Input equipment, handwriting characteristic extraction Module, person's handwriting SVG models, handwriting model training module, dictionary association algoritic module, interface information control module, display device, Wherein：Input equipment be connected with handwriting characteristic extraction module and transmit user input person's handwriting a series of plane coordinates, person's handwriting Characteristic extracting module and handwriting model training module are connected with person's handwriting SVG models and input person's handwriting characteristic information and input respectively As a result feedback information, person's handwriting SVG models associate algoritic module and interface information control module with dictionary and are connected and export pen respectively The result of mark identification, dictionary association algoritic module are connected with interface information control module and export what is filtered out under association's pattern Most possible word, interface information control module are connected with display device and export animation information, and user is according to input results Feedback information is passed to handwriting model training module by correctness, for training new person's handwriting SVG models, is known so as to improve constantly Other accuracy rate.
- 2. system according to claim 1, it is characterized in that, under association's input pattern, person's handwriting SVG models are associated with dictionary Algoritic module be connected and transmit person's handwriting identification result, it is possible to letter and its probability；Under non-association's pattern, person's handwriting SVG Model is directly connected with interface information control module, and the result for transmitting person's handwriting identification selects for user, according to the big I of probability The quantity of result is selected to be no more than 3.
- 3. system according to claim 1, it is characterized in that, described person's handwriting SVG models, by on-line study mechanism to The handwriting trace at family and correct recognition result carry out learning classification, to be optimized to SVG models, so as to further improve word Accord with the accuracy rate of classification.
- 4. a kind of text entry method of the system according to any of the above-described claim, it is characterised in that comprise the following steps： Rocking bar in being moved first by using rocking bar is touched or leaves border, rocking bar returns to center and reverse three kinds of states on border To split track and split characteristic vector pickup track characteristic according to curvature threshold, and it is trained to the book for Real time identification user The SVG models of track are write, then character classification is carried out to the same rocking bar input trajectory for carrying out feature extraction using SVG models and sentenced It is disconnected, and export a series of possible letters and its probability.
- 5. according to the method for claim 4, it is characterized in that, switch between association's input pattern and non-association's input pattern, Under association's input pattern, the character classification judged result of all tracks inputted will calculate with the word in dictionary combines Probability provides probability highest word, and is constantly updated with input is continued；Under non-association's input pattern, probability is higher than threshold value And character judged result of the quantity no more than three will be supplied to user to be selected.
- 6. the method according to claim 4 or 5, it is characterized in that, described method, specifically include following steps：Step 1, collection initial user letter track sample characteristics simultaneously train SVG learning models, specifically include：1.1 read the plane coordinates of alphabetical handwriting trace；1.2 moved according to rocking bar in rocking bar touching or leave border, rocking bar returns to center and reverse three kinds of states on border Judge whether to add new stroke, each stroke is made up of a series of coordinate points；1.3 screen according to sampled point quantity to stroke, remove noise；Each stroke in 1.4 tracks is split further according to certain curvature threshold to be similar to the characteristic vector of straightway；1.5 from the multinomial feature of each characteristic vector pickup, including distance, angle, absolute position, absolute angle, absolute distance and Skew；1.6 train SVG models according to track characteristic set；Step 2, a rocking bar input trajectory is identified, specifically included：A 2.1 rocking bar input trajectory coordinate datas of collection；2.2 extraction track characteristics；2.3 import SVG learning models；2.4 output recognition results, comprising multiple letters and its probability, when present mode is associates input pattern, into step 3, otherwise into step 4；Step 3, according to track identification result dictionary is screened, specifically included：3.1 remove the word that possibility is lost in dictionary according to this recognition result；3.2 judge now whether dictionary is empty；If 3.3 dictionaries, which are sky, provides prompting, if not being superimposed impossible letter in alphabetical group for empty remove in input field；Letter is inputted in step 4, selection candidate list, is specifically included：4.1 screen according to probability threshold values to recognition result, only retain the letter that probability is more than threshold values；4.2 show probability highest letter by default, and in addition first three recognition result of probability is as candidate list；If there is no target alphabetical in 4.3 acquiescences or candidate list, delete current results and re-enter, otherwise continue in next step；The selected letter of 4.4 users；The selected letter of 4.5 input fields display input；4.6 are successfully entered after letter according to alphabetical and corresponding input trajectory renewal SVG models；Step 5, selection word continue input letter, specifically include：5.1 calculate joint probability using HMM algorithms according to the possible result of each letter and the word frequency of probability and word in word；5.2 are arranged word in dictionary as candidate list by probability from big to small, and to picture default word and word candidates table It is updated；If 5.4 select word from candidate list, the word is inputted；If do not selected, continue to input alphabetical track, repeat Step 1~4, until selected word；5.4 words input successfully after according to alphabetical and corresponding input trajectory update SVG models.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510305823.XA CN104898855B (en) | 2015-06-05 | 2015-06-05 | Based on text input system and method with rocking bar equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510305823.XA CN104898855B (en) | 2015-06-05 | 2015-06-05 | Based on text input system and method with rocking bar equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104898855A CN104898855A (en) | 2015-09-09 |
| CN104898855B true CN104898855B (en) | 2017-12-19 |
Family
ID=54031559
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510305823.XA Expired - Fee Related CN104898855B (en) | 2015-06-05 | 2015-06-05 | Based on text input system and method with rocking bar equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104898855B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109074095B (en) * | 2017-12-26 | 2022-04-01 | 深圳市大疆创新科技有限公司 | Flight trajectory original path rehearsal method and aircraft |
| CN109126125A (en) * | 2018-07-03 | 2019-01-04 | 西交利物浦大学 | Text entry method based on double rocking lever controller under reality environment |
| CN109758763B (en) * | 2019-01-07 | 2022-05-13 | 网易（杭州）网络有限公司 | Character input method and device based on game paddle |
| CN117409422A (en) * | 2023-12-15 | 2024-01-16 | 吉林大学 | Oracle retrieval method based on handwriting input |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5841373A (en) * | 1994-07-05 | 1998-11-24 | Mason; Donny L. | Simplified alphabetic communicator for handicapped or infirmed persons |
| CN1900895A (en) * | 2006-07-14 | 2007-01-24 | 杜力 | Method for input character, letter or number using rocking rod key |
| CN101256443A (en) * | 2007-02-27 | 2008-09-03 | 黄金富 | Input device and corresponding input method being suitable for game machine and handhold equipment |
| WO2011057586A1 (en) * | 2009-11-16 | 2011-05-19 | 北京搜狗科技发展有限公司 | Method and apparatus for character input using direction operation |
| CN102156890A (en) * | 2011-04-25 | 2011-08-17 | 汉王科技股份有限公司 | Method and device for stroke segmentation of handwritten formula |
| CN102810008A (en) * | 2012-05-16 | 2012-12-05 | 北京捷通华声语音技术有限公司 | Air input system, method and air input acquisition equipment |
| CN103425269A (en) * | 2010-11-16 | 2013-12-04 | 北京搜狗科技发展有限公司 | Method and device for utilizing directional operations to input characters |
-
2015
- 2015-06-05 CN CN201510305823.XA patent/CN104898855B/en not_active Expired - Fee Related
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5841373A (en) * | 1994-07-05 | 1998-11-24 | Mason; Donny L. | Simplified alphabetic communicator for handicapped or infirmed persons |
| CN1900895A (en) * | 2006-07-14 | 2007-01-24 | 杜力 | Method for input character, letter or number using rocking rod key |
| CN101256443A (en) * | 2007-02-27 | 2008-09-03 | 黄金富 | Input device and corresponding input method being suitable for game machine and handhold equipment |
| WO2011057586A1 (en) * | 2009-11-16 | 2011-05-19 | 北京搜狗科技发展有限公司 | Method and apparatus for character input using direction operation |
| CN103425269A (en) * | 2010-11-16 | 2013-12-04 | 北京搜狗科技发展有限公司 | Method and device for utilizing directional operations to input characters |
| CN102156890A (en) * | 2011-04-25 | 2011-08-17 | 汉王科技股份有限公司 | Method and device for stroke segmentation of handwritten formula |
| CN102810008A (en) * | 2012-05-16 | 2012-12-05 | 北京捷通华声语音技术有限公司 | Air input system, method and air input acquisition equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104898855A (en) | 2015-09-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101620480B (en) | Method for realizing handwritten stroke input on touch screen | |
| CN102193736B (en) | Input method and system supporting multimode automatic switching | |
| CN104898855B (en) | Based on text input system and method with rocking bar equipment | |
| CN103576886B (en) | A kind of numbers and double spelling syllables double-stoke input method and its keyboard plan | |
| EP3084580A1 (en) | User interface for overlapping handwritten text input | |
| CN104704452A (en) | A split virtual keyboard on a mobile computing device | |
| CN102184074B (en) | Keyboard, automatic hand input switching method and device and mobile terminal | |
| CN102880302A (en) | Word identification method, device and system on basis of multi-word continuous input | |
| US8713464B2 (en) | System and method for text input with a multi-touch screen | |
| CA2749916A1 (en) | Data entry system | |
| CN1777858A (en) | Unambiguous text input method for touch screens and reduced keyboard systems | |
| CN103246437A (en) | Approach for switching input methods and terminal | |
| Kumar et al. | Online Devanagari isolated character recognition for the iPhone using Hidden Markov Models | |
| CN102147706A (en) | Method for inputting full spellings of Chinese character in touching and sliding manner | |
| CN101825955A (en) | Eight-final pinyin input method | |
| CN102929518A (en) | Virtual keyboard and information inputting method based on virtual keyboard of touch screen | |
| CN102779003A (en) | Shape input method of Chinese characters | |
| CN104598045B (en) | Five checker input method | |
| JPH07146918A (en) | Handwritten character recognizing device | |
| TWI676932B (en) | Input method and system of electronic device | |
| CN105183183A (en) | Phonetic and shape combination quick input coding and optimization method for square Miao writing | |
| CN104834392B (en) | A kind of Chinese character input method of stroke dynamic group word | |
| CN105242795A (en) | Method for inputting English letters by azimuth gesture | |
| CN106339726A (en) | Method and device for handwriting recognition | |
| CN106201003A (en) | A kind of dummy keyboard based on touch panel device and input method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20171219 Termination date: 20200605 |