CN1224889C - Chinese character input method and system by using said method - Google Patents
Chinese character input method and system by using said method Download PDFInfo
- Publication number
- CN1224889C CN1224889C CN 02126464 CN02126464A CN1224889C CN 1224889 C CN1224889 C CN 1224889C CN 02126464 CN02126464 CN 02126464 CN 02126464 A CN02126464 A CN 02126464A CN 1224889 C CN1224889 C CN 1224889C
- Authority
- CN
- China
- Prior art keywords
- input
- feature
- font
- chinese character
- character
- 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
The present invention relates to an input method and an input system thereof for single Chinese characters. The system comprise a first input device, a second input device, a phonetic recognition device, a font recognition device, a character library device, a Chinese character feature processing device and a display device. The method comprises the steps that the first input device inputs the phonetic features of a Chinese character, and the second input device inputs N font features of the same Chinese character; the phonetic recognition device and the font recognition device respectively identify the phonetic features and the font features; the Chinese character feature processing device combines the phonetic features with the font features of the same Chinese character, selects one candidate Chinese character set which accords with the phonetic features and the N font features from the character library, and outputs and displays the candidate Chinese character set.
Description
Invention field
The present invention relates to the Chinese character input, in particular to the method for single Chinese character and the system of this method of use are imported in the phonetic feature and the combination of font feature of Chinese character.
Background of invention
Along with computing machine use universal day by day, computing machine no longer is simple science computational tool already, but becomes the aid in people's daily life and the work.Therefore how to improve the interactive efficiency of people and computing machine, reduction user's mutual burden, become an important problems.And the input of information is the most key in this a problem aspect.
Chinese character input method at present popular on desktop computer can be divided into two kinds substantially, and one is to use the method (as: Microsoft's input method) of phonetic, and two are to use the method (as: the Five-stroke Method) of Hanzi structure, and their input tool all is a keyboard.Along with the continuous development of human-computer interaction technology, new Chinese character input method is constantly emerged in large numbers, and wherein writing pencil input and phonetic entry are to it seems the most promising input method at present, and this is because hand-written and voice are the most natural communication meanses of people.Simultaneously, along with a large amount of appearance and the use of handheld mobile devices such as palm PC, keyboard can't be brought into play its effect greatly because of volume in the environment that moves, and this provides bigger usage space for handwriting input and phonetic entry.Background technology related to the present invention comprises:
1. speech recognition technology
Speech recognition technology is treated to research object with voice signal, relates to numerous areas such as linguistics, computer science, signal Processing, physiology, psychology, is the important branch of pattern-recognition.Be the beginning period of The Research of Speech Recognition work the fifties, the system that discerns ten numerals that its sign Bell Laboratory is succeeded in developing.The sixties, computing machine is widely used in the research work of speech recognition, and dynamic programming and linear prediction analysis technology are the important achievement in this period.The seventies, the research of speech recognition has obtained breakthrough, in reality is explored, is developed success based on the isolated sound recognition system of specific people of linear prediction cepstrum and dynamic time warping technology, on theoretical method, vector quantization and hidden Markov model theory have been proposed then.The eighties, the research work of speech recognition further deeply.Its sign is the successful Application of artificial neural network in speech recognition.The nineties, with fast development of computer technology, speech recognition moves towards practical from research, and its achievement in research has reached quite high level.
The Via Voice speech recognition software that IBM Corporation releases is a successful examples of speech recognition.It is an identification software of supporting continuous speech input and discontinuous speech input, and the engine of a cover voice identification can also be provided simultaneously.
But, speech recognition at present also has its weak point, and one of them important problem is: because the phonetically similar word of Chinese character is numerous, therefore when the user carries out the individual character phonetic entry, how filtering out in phonetically similar word and want the Chinese character imported, is the problem that needs solve.In the Via of IBM Voice software, system provides the highest word of frequency of utilization of oneself acquiescence, if what the user will import is not this word, then can all phonetically similar words of this word be listed by specific operational order system, carry out secondary by the user and select.Therefore, the problem of this method is: for the numerous individual character of some phonetically similar words, selection course is a very big burden to the user.
2. handwritten Kanji recognition technology
Along with the development of pattern-recognition and the artificial intelligence of computing machine, on the basis of English and digit recognition, Chinese character recognition technology grew up the seventies in 20th century, and had experienced the process that is recognized hand script Chinese input equipment identification by printed Chinese character.
Printed Chinese character is discerned early start in Japan, and has carried out open demonstration in the printed Chinese character recognition device to its development in 1980, and this device can be discerned 2000 of Chinese characters, and recognition speed was 100 word/seconds, and recognition accuracy at that time reaches 98.4%.The Chinese character hand-written identification of China starts from eighties of last century seventies, and the research direction of beginning also concentrates on the handwriting recognition of block letter, has obtained the phasic results of block letter handwriting recognition research in earlier stage in the nineties.
Universal day by day along with science and technology development and computing machine, the eighties in 20th century, online Chinese character hand-written identification has been subjected to increasing attention.The Chinese Character Recognition of hand script Chinese input equipment is the system of a real-time, interactive, and the user carries out handwriting input by pen and corresponding input equipment on one side, on one side user's input is gathered and discerned to computer system, and recognition result is returned to the user.Aspect the research of hand script Chinese input equipment identification, Han Wang Technology Co., Ltd has obtained very proud achievement, and the Chinese king of its an exploitation Chinese character hand script Chinese input equipment recognition system can be discerned various fonts such as traditional font, simplified, running hand, rapid style of writing.
From the theoretical method of Chinese Character Recognition, the method for Chinese Character Recognition is divided into several big classes such as statistical recognition, structure identification and neural net method basically.What a large amount of hand script Chinese input equipment recognition systems adopted all is structural recognition method.So-called structural recognition method, its starting point is the composition structure of Chinese character, on the formation of Chinese character, Chinese character is to be made of stroke (point, horizontal, vertical, left-falling stroke, right-falling stroke etc.), radical, radicals by which characters are arranged in traditional Chinese dictionaries, by a complicated Chinese mode being decomposed into simple subpattern until the basic model element, subpattern is judged and carried out the matching algorithm computing based on symbolic operation, reach identification to complex patterns.The advantage of structure method of identification is that the ability of differentiation similar character is strong, and shortcoming is a poor anti jamming capability.Statistical recognition method be Chinese character is seen as a whole, its all feature from this on the whole through big quantitative statistics and obtaining, then according to the judgement of classifying of the determined decision function of certain criterion.The characteristics of statistical recognition are strong interference immunities, shortcoming be the segmentation ability a little less than.
The general process of handwriting recognition comprises several main stages of data acquisition, pre-service, normalization, feature extraction, characteristic matching and output character code.Data acquisition generally is that the user utilizes an equipment to realize by input equipments such as handwriting pad or touch-screens, the stroke that the user writes on these equipment is stored down with the form that is similar to polar plot, system is by the pen of lifting to character and picture, start to write, after information such as time relationship on the person's handwriting between the locus of each pixel and the pen section are carried out pre-service and normalized, with certain Rule Extraction characteristic information, by recognition device characteristic information and intrasystem identification storehouse are compared and contrast afterwards, discerned, finally transformed into the employed character code of computing machine.
What most on-line handwritten Chinese character identification was adopted is the method for structured mode identification.Concrete can be divided into two classes again: multistage identifying schemes and single-stage identifying schemes.In multistage identifying schemes, recognizer always identifies each independently stroke earlier, then closes to tie up to according to the position between the stroke to determine whole Chinese character.In the single-stage identifying schemes, as a whole identification done in the whole Chinese character after system directly finishes according to input.
Handwriting recognition is unusual maturation on general common computer at present, but the effect on palm PC but is not an ideal very.This is because the storage capacity and the computing power of handheld mobile device are limited, makes the size in identification storehouse and speed, the accuracy of coupling all be subjected to influence in various degree.
3. based on the Chinese character input method of the order of strokes observed in calligraphy
In common in the market mobile phone, all provide " structure input ".Its main thought is: with the specific stroke of button representative, write down the order of strokes observed in calligraphy information of each Chinese character in character library, button once is equivalent to import a stroke, and system can show all Chinese characters that meet order of strokes.
This method requires the number of strokes of user's input many, and promptly button is easy to increase the sense of fatigue of user's input often like this.
4. with voice and the hand-written Chinese character input method that combines
The Chinese patent notification number is 1112252 patent, has invented a kind of with voice and the hand-written computer input tool that combines.This invention combines input Chinese character of the handwriting pad pen type under same operating system and phonetic entry Chinese character with polling mode, these two kinds of Chinese character input tools just need not be switched can call the input Chinese character mutually.This invention has improved the automaticity of Comnputer Chinese character input, and makes and manyly can't also can use Chinese character inputting with the personage of keyboard as the Chinese character input tool because of a variety of causes.
In addition, Chinese patent authorization notification number is in 2198630 the patent, the inventor has invented the digital pen of a kind of voice and hand-written pair of identification, it is to install the microphone that can be used as phonetic entry additional in an appropriate location that can cooperate the numerical digit plate to make the digital pen handwriting of hand-written signal input, has difunctional (phonetic entry and become, handwriting input) digital pen, this novel digital pen needs with phonetic entry when wrong auxiliary doing the handwriting input identification, directly with pen with regard to mouth do phonetic entry or phonetic entry when wrong directly with hand-written correction, so mutual correction can obtain fast, save space, the combined effect of binary channels identification easily, promptly it can improve the efficient of the two identification systems of computer.
Above-mentioned two inventions all are with voice and hand-written combining, and make both reach complementary effect, make people's input more convenient, fast, nature.But, in the superincumbent invention,, remain and utilize speech recognition or handwriting recognition technology respectively specific to the input of each Chinese character, therefore with regard to the input efficiency of single Chinese character, do not improve.
The objective of the invention is to, a kind of Chinese character input system of importing the method for Chinese character and using this method is provided, wherein, when each independent Chinese character of input, voice and two kinds of information of hand-written stroke can be combined, dwindle the scope of candidate, improve the input speed of individual character, thereby improve the input speed of literal integral body, thereby reduce user's input burden.
Summary of the invention
This purpose is achieved by Chinese character input method of the present invention.
According to the present invention, a kind of method of importing Chinese character on information equipment is provided, described equipment has an input system, comprise first input media, second input media, speech recognition equipment, font recognition device, character library device, Hanzi features treating apparatus and display device, described method comprises:
Phonetic feature by first input media input Chinese character also outputs to speech recognition equipment and/or second input media is imported N font feature of same Chinese character and outputed to the yi word pattern recognition device;
Utilize described speech recognition equipment and font recognition device to discerning, and recognition result is outputed to the Hanzi features treating apparatus respectively from the phonetic feature of first input media with from the individual font feature of N (N is the natural number greater than 1) of second input media respectively;
Utilize the Hanzi features treating apparatus that the output result from speech recognition equipment and font recognition device is respectively handled, phonetic feature and N font feature of same Chinese character are combined, from described character library, select and both meet candidate Chinese character set that described phonetic feature also meets described N font feature, and with its output;
To be integrated on the display device from the described candidate Chinese character of Hanzi features treating apparatus and show.
Preferably, adopt the preferential scheme of voice channel according to Chinese character input method of the present invention.The Hanzi features treating apparatus is at first according to extracting the Chinese Character Set that meets this phonetic feature from phonetic feature first input media and that discern through speech recognition equipment from described character library, then according to from described voice Chinese Character Set, selecting the character set that also meets described N font feature, and this had both been met candidate Chinese character that described phonetic feature also meets described N font feature gather and output to display device from second input media and through N font feature of the same Chinese character of font recognition device identification.Voice channel is the main channel under this scheme, and the effect of font information is the scope of dwindling voice identification result.
Here, described information equipment comprises equipment such as palm PC, mobile phone, notebook computer, electronic notebook, desktop computer.Described first input media comprises voice-input devices such as microphone.Second input media comprises pointing apparatus such as writing pencil/handwriting pad, mouse, keyboard, touch-screen, preferably uses writing pencil/handwriting pad.Described phonetic feature refers to that the user treats input sound wave that individual character sent or corresponding to the sound signal of waiting to import individual character.Described font feature refers to form that the element (as stroke, radical) of Chinese character itself had together with the information such as sequential write, locus or orientation of this element at the place literal, the corresponding described feature of element with described information.For example, in the order of strokes scheme, " king " word has four font features, be respectively the first stroke " horizontal stroke ", second " horizontal stroke ", the 3rd " erecting " and the 4th " horizontal stroke ".
Among the present invention, preferably use the order of strokes feature, but do not get rid of other schemes.In addition, above-mentioned M is the number of the font feature that each Chinese character had in the described character library, among some embodiment in the present invention, and M=3, certainly, also desirable other positive integer of M.
According to the present invention, in each single Chinese character of input, all the feature of its sound and the feature of type are combined according to input system of the present invention, utilize two passages of voice and font to carry out the Chinese character input simultaneously.That is to say that the user can say while writing, and finishes once input, has significantly reduced the quantity of candidate, improves user's input speed in Chinese character of input.This input method is fit to not have association, the more weak Chinese character of semantic relation therebetween, as the input of name, place name etc., is particularly suitable for using on miniature portable information equipment such as mobile phone, palm PC.
According to the present invention, the input system of using input method of the present invention also is provided, comprising:
First input media is used to import the phonetic feature that desire is imported Chinese character, and with its output;
Second input media is used to import the individual font feature of N (N is the natural number more than or equal to 1) that described desire is imported Chinese character, and with its output;
Speech recognition equipment is used to receive from the phonetic feature of first input media and discerns, and the result is exported;
The font recognition device is used to receive from the font feature of second input media and discerns, and the result is exported;
Character library device, file be the Chinese character of predicate sound feature and/or font feature to some extent;
The Hanzi features treating apparatus, be used to receive respectively from the output result of speech recognition equipment and font recognition device and handle, (N=1 is to M with the phonetic feature of same Chinese character and N, N and M are positive integer) individual font feature combines to select from described character library and both meets candidate Chinese character set that described phonetic feature also meets described N font feature, and with its output;
Display device is used to show the result of Hanzi features treating apparatus,
Preferably, adopt the preferential scheme of voice channel according to input system of the present invention.The Hanzi features treating apparatus is at first according to extracting the Chinese Character Set that meets this phonetic feature from phonetic feature first input media and that discern through speech recognition equipment from described character library, then according to from described Chinese Character Set, selecting the character set that also meets described N font feature, and this had both been met candidate Chinese character that described phonetic feature also meets described N font feature gather and output to display device from second input media and through N font feature of font recognition device identification.Voice channel is the main channel under this scheme, and the effect of font information is the scope of dwindling voice identification result.
Below in conjunction with accompanying drawing embodiments of the invention are described in detail, advantage of the present invention and characteristics will be more obvious.
Brief description
Fig. 1 is the structured flowchart that has the input system of Chinese character input method of the present invention.
Fig. 2 is system according to the present invention first processing procedure during the input voice under the voice channel preference strategy.
Fig. 3 is system according to the present invention first processing procedure during the input stroke under the voice channel preference strategy.
Fig. 4 is system according to the present invention first processing procedure during input font feature under font feature preference strategy.
The processing procedure of Fig. 5 another one embodiment of the present invention, wherein the character library device comprises at least two character libraries.
Preferred embodiment describes in detail
Fig. 1 is the synoptic diagram of the functional block that comprises of input system of the present invention.Input system according to the present invention comprises: first input media 1 is used to import the phonetic feature that desire is imported Chinese character, and it is outputed on the speech recognition equipment 3; Second input media 2 is used to import the individual font feature of N (N=1 is to M, and N, M are positive integer) that described desire is imported Chinese character, and it is outputed on the font recognition device 4; Speech recognition equipment 3 is used for discerning from the phonetic feature of first input media, and the result is outputed on the Hanzi features treating apparatus 5; Font recognition device 4 is used for discerning from the font feature of second input media 2, and the result is outputed on the Hanzi features treating apparatus 5; Character library device 6 stores a large amount of with the phonetic feature and/or the Chinese character of the individual font feature of M (M is a predetermined positive integer, for example 3) at least; Hanzi features treating apparatus 5, be used for the output result from speech recognition equipment 3 and font recognition device 4 is respectively integrated, from the character library of described character library device, select a Chinese Character Set that comprises desire input Chinese character according to described phonetic feature and/or N font feature, and it is outputed to display device; Display device 7 is used to show the result of Hanzi features treating apparatus.Wherein, although be discrete on the function, second input media can physically can be an one with display device.
In input system according to the present invention, the user can carry out the input of single Chinese character by following several modes:
(1) phonetic feature is imported separately.The user by first input media for example microphone carry out phonetic entry, then in numerous phonetically similar words that system provides, pick out one that wants to import.For example, when the user by voice channel input " I (wo) ", all unisonance Chinese characters " I irrigate hold the nest that crouches revolve the wet tent of snail whirlpool Laos Japan " of " I " will be selected by system.Then the user selects single Chinese character " I " in these phonetically similar words, finishes input.In this case, be similar with prior art.
(2) the font feature is imported separately.The user in handwriting area, utilize the font feature for example adopt order of strokes observed in calligraphy scheme by second input media for example writing pencil/handwriting pad carry out handwriting input, behind the certain hour of system after the user stops to import, call the device of handwritten form identification, this input of user is discerned, the result of identification is returned to the user.In this case, also be similar with prior art.
(3) the combination input of phonetic feature and font feature.According to the present invention, the user can combine " writing while saying " with above-mentioned two kinds of inputs.The user can carry out phonetic entry and pen input synchronously, and system can integrate the information of two passages, finds the Chinese character that meets two channel informations simultaneously.For example, the user is in input " wo ", if also imported stroke " left-falling stroke " by pen, then system can be that " wo " and all the first strokes are that the Chinese character of casting aside provides i.e. " my Japan " with voice, if the user has imported one " erecting " immediately again, then system then can choose " Japan " automatically.Can significantly reduce the quantity of candidate like this, improve user's input speed.
In one embodiment of the invention, what speech recognition was adopted is the individual character recognition engine of IBM Voice, and recognition device will return the highest one of frequency of utilization in the phonetically similar word of user's pronunciation; Second input media 2 adopts writing pencil/handwriting pad equipment, and requirement can identify independent stroke in native system, can identify whole Chinese character again.When discerning independent stroke, adopt method of discrimination based on stroke direction and direction variation; When the whole Chinese character of identification, system adopts present popular algorithm based on template matches to discern.
In the character library in the present embodiment, each Chinese character all has phonetic feature and font feature simultaneously.For improving retrieval rate, adopted following scheme: character library is made up of two parts.First is the index part of phonetically similar word, in this part, represents word for one of each Chinese character pronunciation that comprises in the in store character library.When speech recognition equipment receives the user's voice input at every turn, all can call identifying, the result of the each identification of this process is a fixing word of this pronunciation.For example, when the user sends " wo " this sound, all can return fixed word " I ", this word is exactly the representative word of " wo " this pronunciation.Represent the character library call number of all Chinese characters of the in store unisonance with it of word.
Second portion is the main part of character library.In main part, Chinese character is organized in together by the information of sound, and promptly the Chinese character of all unisonances is stored in together, and the index of device is kept at the representative word place of index part.In addition, in order to write down the font feature of each Chinese character, the coding of M stroke feature in the present embodiment, was got M=3 before each Chinese character back was all in store.Certainly, M also can get other suitable positive integer as required.
In this scheme, voice channel is the main channel, and the effect of gesticulating passage is the scope of dwindling voice identification result.We are divided into two classes to user's input mode: input voice and input stroke earlier earlier.
Below, in conjunction with Fig. 1, we are described Chinese character input processing procedure equipment according to the present invention in Fig. 2.
A. import voice earlier
See Fig. 2.Processing procedure when Fig. 2 is first input of equipment according to the present invention voice under the voice channel preference strategy.
At first by first input media, 1 input voice (step 200), this phonetic feature is discerned (step 201) through phonetic feature recognition device 3, produces a specific recognition result.With this recognition result as representing word in character library device 6, to retrieve (step 202), the unisonance character set that obtains and their stroke information are kept in the character buffering (step 203), and it is presented at selects for the user in the specific region of display device 7 or system selects desire input Chinese character (step 204) automatically.After this process finished, system closed voice channel automatically, and having chosen desire input Chinese character or pressure up to the user, that current input is set is invalid.
When the user carries out handwriting input (step 205) by second input media 2, font recognition device 4 will identify the user and import stroke, export specific result and will give Hanzi features treating apparatus 5, and start clock, and the stroke number setting adds 1 (step 206).Hanzi features treating apparatus 5 will be according to from the particular result of the output of font recognition device 4 the stroke information of the character set that extracts according to phonetic feature in the character buffering being mated.The word that meets the stroke information that the user imports is retained, and incongruent word is removed from the character buffering, thereby dwindles the character set scope, produces candidate Chinese character set (step 209).One of the every input of user, system all can carry out the described process of a deuterzooid section, and the number of strokes N that appears at candidate Chinese character set or input up to desire input Chinese character reaches M.When stroke number reaches M, system will close hand-written passage (step 207) automatically.
The user clicks desire input Chinese character by an equipment, and system finishes this input process with the input of this Chinese character as this.If have only a word in the character buffering, system can choose this word automatically, then finishes this input process (step 204).
After input process finished, system carried out the replacement of resource.Comprise: an opening voice passage and a passage again, the input of removing user's handwriting area, and remove used all temporary resources and variable in the identifying.
B. import by pen earlier
See Fig. 3.Fig. 3 is equipment according to the present invention first processing procedure during the input stroke under the voice channel preference strategy.
When the user by second input media 2, for example, writing pencil/handwriting pad equipment is finished (step 300) after the input of the first stroke, the stroke number setting adds 1, and starts clock (step 301).The value representation of clock: the maximum time between continuous two inputs of same word at interval.In case surpass this time interval, this end of input is thought by system, only according to handwritten form whole Chinese character is discerned (step 305).
According to the experiment that 20 subjects are carried out,, then generally can when second of input or the 3rd, carry out phonetic entry if the user wishes with voice pen input to be assisted.In view of the above, default: if user's input has surpassed 3 (step 302), then system closing voice channel (step 303) is waited for when clock is overtime, according to handwritten form whole Chinese character is discerned.
Importing preceding 3 and timer the user does not have under the overtime situation, and font feature identification device 4 can be discerned (step 306) to the stroke of user's input.Whether have phonetic entry arrive (step 307), if do not have, then stroke is kept at (step 308) among the stroke buffering if then detecting.In case the user has carried out importing (step 310) by first input media, voice channel will become the main channel, 3 pairs of phonetic features of speech recognition equipment are discerned (step 311), character library is retrieved (step 312), draw the character set (step 313) that meets phonetic feature, then the character set in the character buffering is screened (step 314) according to the stroke information of preserving in the stroke buffering.Matching process when the process of this coupling screening is imported voice with elder generation is identical, and at this moment the state of system just is equivalent to first voice, and the back is hand-written.If user's input stroke is less than 3, then the user can continue to import by pen, and the character set in the character buffering is further screened (step 309), and the stroke counting is postponed.Finally, behind selected desire input Chinese character, with its demonstration, and replacement system resource (step 315).
In the superincumbent identifying, in case mistake has taken place in the identification of certain passage, the user can finish this input process by forced system, and system can reset to resource automatically, gets back to the preceding state of this input beginning.
Under the preferential strategy of voice channel, if the stroke input information arrives first, system will be kept at it stroke buffering, when phonetic entry by the time arrives, then, obtain character set, then screen according to the stroke information in the stroke buffering at once according to phonetic feature retrieval character library.
In another embodiment, adopt the main scheme that is input as earlier, both if voice arrive first, then voice channel was main, if the font feature arrives first, then the font passage is main.The former has comprised that in the foregoing embodiments we describe the latter now.If the font feature arrives first as an input information, then according to order of strokes observed in calligraphy information retrieval character library, obtain a character set, when the user imports by voice, this character set is screened with phonetic feature.
Particularly, as shown in Figure 4, when the user imports by second input media 2, finish (step 400) after the input, the stroke number setting adds 1, and starts clock (step 401).If the user is at the N=M pen of having imported a Chinese character, for example after 3 (step 402), voice channel does not still detect voice input signal, and then system will close voice channel (step 403).After clock is overtime (step 404), only discern (step 405) according to handwritten form.
Import N stroke and timer does not have under the overtime situation the user, system can carry out independent identification (step 406) to the stroke of user's input, retrieval qualified character set (step 407) in character library leaves the result in the character buffering (step 408) in then.At this moment if received the phonetic entry (step 409) of passing through first input media, speech recognition equipment is discerned (step 410) to the voice of input, screen character set in the character buffering according to the result of identification, find out the set of the Chinese character that meets two kinds of features, be retained in (step 411) in the buffer zone.If when the stroke number N＜M of user's input and/or described character set do not comprise desire input Chinese character, can continue to import stroke and further described character set be screened.When the word of having selected the desire input or candidate had only one, system was presented at this Chinese character the viewing area and resets all resources automatically.
See Fig. 5.Fig. 5 is another one embodiment of the present invention, and wherein character library device 4 comprises at least two character libraries, and one comprises the phonetic feature character library that is used for the phonetic feature retrieval, and each Chinese character does not have stroke information coding.And the Chinese character in the another one character library all has the stroke information characteristics coding of M position (being 3) here.This system is provided with unique character buffering.
No matter two passages that received user's information, the recognition device that all calls is separately discerned (step 500,501 to this input, 505,506), then from character library device retrieval character library (step 502,510) two character sets that, met the input information feature description respectively.Then, two character sets are carried out cap, the result is kept in the character buffering (step 503).Enough hour of Chinese total number in buffering, the user clicks Chinese character to be selected by an equipment, and selected this Chinese character finishes this input process as this input.If have only a word in the character buffering, system can choose this word automatically, then finishes this input process (step 504).
In the superincumbent discussion, if some or two identification error has taken place simultaneously in two passages, cause occuring simultaneously for empty, then the identification set according to voice signal is as the criterion (certainly, also can ignore voice signal, be as the criterion according to the stroke of handwriting input).
In another one embodiment of the present invention, on the basis with top dual mode, replace passing through the handwriting input of writing pencil/handwriting pad equipment with keyboard.A series of button promptly is provided on keyboard, and each button is represented a font characteristic element, stroke feature for example, totally 5---horizontal, vertical, cast aside, point and other.Can certainly adopt other font structure feature scheme, as the Five-stroke Method.The user is carrying out stroke information when input, can be not directly hand-written by an equipment, but button carries out, such benefit is to make the discrimination of stroke information increase.
Certainly, also can have hand-written simultaneously and the button dual mode, which the user uses decide according to the personal like fully.
In input during Chinese character, all the characteristics of the sound of word and the characteristics of type are combined according to input system of the present invention, utilize two passages of voice and font to carry out the Chinese character input simultaneously.That is to say that the user can say while writing, and finishes once input, thereby significantly reduces the quantity of candidate, has improved user's input speed in Chinese character of input.This input method is suitable for importing the Chinese character of association of mutual nothing and semantic relation, as name etc., is particularly suitable for using on miniature portable information equipment such as mobile phone, palm PC.
In above-mentioned detailed description, the embodiment of reference specific exemplary wherein is illustrated method and apparatus of the present invention.Yet, clearly, can carry out various modifications, combination and variation to this and not depart from the present invention's scope widely.Present technique explanation is corresponding with accompanying drawing to be considered to illustrative, but not determinate.
Claims (24)
- One kind on information equipment the input Chinese character method, described information equipment comprises an input system, described input system comprises first input media, second input media, speech recognition equipment, font recognition device, character library device, Hanzi features treating apparatus and display device, and described method comprises the steps:Output to speech recognition equipment and second input media input font feature from the phonetic feature of first input media input Chinese character and with it and it is outputed on the font recognition device;Utilize described speech recognition equipment and font recognition device to discerning from the phonetic feature of first input media with from the font feature of second input media respectively, and the result is outputed to respectively on the Hanzi features treating apparatus;Utilize the Hanzi features treating apparatus that the output result from speech recognition equipment and font recognition device is respectively handled, from described character library, select a candidate Chinese character set according to described phonetic feature and font feature, and it is outputed on the display device;On display device, show described candidate Chinese character set from the Hanzi features treating apparatus;It is characterized in that, that described Hanzi features treating apparatus is imported respectively by input of described first input media and described second input media and combine from described character library, to select through the phonetic feature of the same Chinese character of speech recognition equipment and the identification of font recognition device and font feature and both meet the candidate Chinese character that described phonetic feature also meets described font feature and gather.
- 2. method as claimed in claim 1, wherein, the element of described font feature is 5, be respectively horizontal, vertical, cast aside, point and other stroke features.
- 3. method as claimed in claim 1 is characterized in that, described character library device comprises a character library at least, and each Chinese character of described character library all has phonetic feature and M font feature, and wherein, M is a positive integer.
- 4. method as claimed in claim 1, wherein, voice channel is set to privileged way.
- 5. method as claimed in claim 4, wherein, if phonetic feature is input earlier, then from described character library, extract the character set that meets this phonetic feature, from described character set, select the candidate Chinese character set that meets described N font feature according to N font feature of input then, and it is outputed on the display device according to described phonetic feature, wherein, N is a positive integer, more than or equal to 1, smaller or equal to M;If the font feature is input earlier, then described font feature is kept in, the input of waiting voice feature, if receive speech input information in the given time, then at first from described character library, extract the character set that meets this phonetic feature according to described phonetic feature, from described character set, select the candidate Chinese character set that meets N font feature according to described temporary font feature and the font feature that may import subsequently then, and it is outputed on the display device.
- 6. method as claimed in claim 1 is characterized in that, the passage of input is the main channel earlier.
- 7. method as claimed in claim 6, it is characterized in that, if phonetic feature is input earlier, then from described character library, extract the character set that meets this phonetic feature according to described phonetic feature, from described Chinese Character Set, select the candidate Chinese character set that also meets described N font feature according to N font feature of input then, and it is outputed on the display device;If the font feature is input earlier, then from described character library, extract the character set that meets this font feature according to the font feature of input, from described Chinese Character Set, select the candidate Chinese character that not only meets phonetic feature but also meet N font feature according to the phonetic feature of input and the font feature that may import subsequently then and gather, and it is outputed on the display device.
- 8. method as claimed in claim 1, wherein, described character library device comprises first character library and second character library at least, and wherein the Chinese character of first character library has phonetic feature, and the Chinese character of second character library has the font feature.
- 9. method as claimed in claim 8, wherein, the described step of Chinese character of selecting from character library comprises: select first set of the Chinese character that meets this phonetic feature from described first character library according to phonetic feature, and from described second character library, select second set of the Chinese character that meets this font feature according to N font feature, cap is carried out in described first set and second set, and will be outputed to described display device as the result's who occurs simultaneously candidate Chinese character set.
- 10. as the method one of among the claim 1-9, wherein, from described character library, select in the step of a candidate Chinese character set, if the included number of words of described candidate Chinese character set is 1, then system selectes this Chinese character automatically as input results, otherwise the Chinese character of selected desire input the described candidate Chinese character set on outputing to described display device.
- 11. the Chinese character input system of using in the information equipment comprises:First input media is used to import the phonetic feature that desire is imported Chinese character, and with its output;Second input media is used to import the font feature that described desire is imported Chinese character, and with its output;Speech recognition equipment is used for receiving from the first input media phonetic feature, and this phonetic feature is discerned, and the result is exported;The font recognition device is used to receive the font feature from second input media, and this phonetic feature is discerned, and the result is exported;The character library device comprises the character library of the Chinese character that has described phonetic feature and/or font feature;The Hanzi features treating apparatus is used for the output result from speech recognition equipment and font recognition device is respectively handled, and selects a candidate Chinese character collection according to described phonetic feature and/or font feature from the character library of described character library device, and with its output;Display device is used to show the result of Hanzi features treating apparatus,It is characterized in that, that described Hanzi features treating apparatus will be imported respectively by input of described first input media and described second input media and combine from described character library, to select through the phonetic feature of the same Chinese character of speech recognition equipment and the identification of font recognition device and N font feature and both meet the candidate Chinese character that described phonetic feature also meets described N font feature and gather, and it is outputed on the display device, wherein, N is more than or equal to 1, smaller or equal to the positive integer of M, M is each word font style characteristic number in the character library, is positive integer.
- 12. the input system as claim 11 is characterized in that, described first input media comprises microphone.
- 13. as the input system of claim 11, wherein, described second input media comprises one of writing pencil/handwriting pad, mouse, keyboard, touch-screen or several combinations.
- 14. as the input system of claim 13, wherein, described keyboard has the button that is set with M font element.
- 15. as the input system of claim 14, wherein, described font element be horizontal, vertical, cast aside, point and other stroke features.
- 16. as the input system of claim 11, wherein, described character library device comprises at least one character library, each Chinese character of described at least one character library all has phonetic feature and font feature.
- 17. as the input system of claim 16, wherein, it is preferential that described input system is set to the voice channel that is made of described first input media.
- 18. input system as claim 17, wherein, if phonetic feature is input earlier, then the Hanzi features treating apparatus is according to extracting the character set that meets this phonetic feature from phonetic feature first input media and that discern through speech recognition equipment from described character library, basis is selected the candidate Chinese character set that meets described N font feature from second input media and through N font feature of font recognition device identification from described Chinese Character Set then, and it is outputed on the display device;If the font feature is input earlier, the Hanzi features treating apparatus is kept in described font feature, the input of waiting voice feature, if receive speech input information in the given time, then the Hanzi features treating apparatus is at first according to extracting the character set that meets this phonetic feature from phonetic feature first input media and that discern through speech recognition equipment from described character library, then according to temporary font feature and may be subsequently from second input media and from described Chinese Character Set, select the candidate Chinese character that also meets described N font feature through the font feature of font recognition device identification and gather, and it is outputed on the display device.
- 19. as the input system of claim 16, wherein, the passage of input is the main channel earlier.
- 20. input system as claim 19, wherein, if phonetic feature is input earlier, then the Hanzi features treating apparatus is according to extracting the character set that meets this phonetic feature from phonetic feature first input media and that discern through speech recognition equipment from described character library, basis is selected the candidate Chinese character set that also meets described N font feature from second input media and through N font feature of font recognition device identification from described Chinese Character Set then, and it is outputed on the display device;If the font feature is input earlier, then the Hanzi features treating apparatus according to from second input media and from described character library, extract the character set that meets this font feature through the font feature of font recognition device identification, then according to from first input media and through the phonetic feature of speech recognition equipment identification and may be subsequently from described Chinese Character Set, select the candidate Chinese character that not only meets described phonetic feature but also meet N font feature and gather, and it is outputed on the display device by the font feature of second input media input.
- 21. as the input system of claim 16, wherein, described character library device comprises two character libraries, wherein first character library has phonetic feature, and second character library has the font feature.
- 22. input system as claim 21, wherein, described Hanzi features treating apparatus basis is selected first set of the Chinese character that meets this phonetic feature from described first character library from the phonetic feature of the identification of speech recognition equipment, and according to second set of from described second character library, selecting the Chinese character that meets this font feature from the font feature of the identification of font recognition device, cap is carried out in described first set and second set, and will be outputed on the described display device as the result's who occurs simultaneously candidate Chinese character set.
- 23. as the input system one of among the claim 11-22, wherein, if the included number of words of described candidate Chinese character set is 1, then system selectes this Chinese character automatically as input results, otherwise utilizes the Chinese character of described second input equipment selected desire input from the shown candidate Chinese character set of described display device.
- 24. as the input system of claim 11, described information equipment is one of mobile phone, palm PC, electronic notebook, notebook computer, Desktop PC.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 02126464 CN1224889C (en) | 2002-07-22 | 2002-07-22 | Chinese character input method and system by using said method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 02126464 CN1224889C (en) | 2002-07-22 | 2002-07-22 | Chinese character input method and system by using said method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1470975A CN1470975A (en) | 2004-01-28 |
| CN1224889C true CN1224889C (en) | 2005-10-26 |
Family
ID=34143309
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 02126464 Expired - Fee Related CN1224889C (en) | 2002-07-22 | 2002-07-22 | Chinese character input method and system by using said method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1224889C (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103186238A (en) * | 2011-12-31 | 2013-07-03 | 中国银联股份有限公司 | Method and system for identifying character input on touch screen |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8447611B2 (en) * | 2006-09-05 | 2013-05-21 | Fortemedia, Inc. | Pen-type voice computer and method thereof |
| TW201229877A (en) * | 2011-01-03 | 2012-07-16 | Compal Electronics Inc | Method for preventing touch keyboard from being mis-touched and electrical apparatus with touch keyboard |
| CN104035551A (en) * | 2013-03-08 | 2014-09-10 | 联想(北京)有限公司 | Input method and electronic device |
| CN103324721B (en) * | 2013-06-26 | 2017-12-26 | 百度在线网络技术（北京）有限公司 | Method, system, browser and the cloud server to be navigated by browser |
| CN104345875A (en) * | 2013-08-07 | 2015-02-11 | 联想(北京)有限公司 | Method for information processing and electronic equipment |
| CN104423543A (en) * | 2013-08-26 | 2015-03-18 | 联想(北京)有限公司 | Information processing method and device |
| CN103631389A (en) * | 2013-11-29 | 2014-03-12 | 乐视致新电子科技（天津）有限公司 | Voice input method and device |
| CN104794485B (en) * | 2015-04-09 | 2018-04-03 | 广东小天才科技有限公司 | A kind of method and device for identifying written word |
| CN107943318A (en) * | 2017-11-10 | 2018-04-20 | 北京小米移动软件有限公司 | Input method and device, electronic equipment and computer-readable recording medium |
-
2002
- 2002-07-22 CN CN 02126464 patent/CN1224889C/en not_active Expired - Fee Related
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103186238A (en) * | 2011-12-31 | 2013-07-03 | 中国银联股份有限公司 | Method and system for identifying character input on touch screen |
| CN103186238B (en) * | 2011-12-31 | 2016-03-09 | 中国银联股份有限公司 | A kind of method and system identifying character input on touch-screen |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1470975A (en) | 2004-01-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1918578B (en) | Handwriting and voice input with automatic correction | |
| TWI552008B (en) | Input processing method and apparatus | |
| CN1224889C (en) | Chinese character input method and system by using said method | |
| CN103092329A (en) | Lip reading technology based lip language input method | |
| CN1512440A (en) | Character input device | |
| CN1607491A (en) | System and method for Chinese input using a joystick | |
| US20040239534A1 (en) | Component-based, adaptive stroke-order system | |
| CN1841280A (en) | Portable keyboard and fingerprint feature information extracting method thereof | |
| CN1237435C (en) | Chinese Character graphic form input device and method | |
| CN101030103A (en) | Method for inputting Chinese words and phrases holographic code by computer and cell phone and its keyboard | |
| CN1140282A (en) | Information processing method and apparatus | |
| CN1141633C (en) | 24-radical sorting encode method for Chinese characters and its keyboard | |
| CN1755589A (en) | Letter inputting method and apparatus based on press-key and speech recognition | |
| CN1991743A (en) | Method and device for voice input method | |
| CN1095118C (en) | Multistage front end processor system | |
| CN1310119C (en) | Flexible keyboard and inputtings style for touch screen | |
| CN1101577C (en) | Speech input memorandum | |
| CN1367420A (en) | Digital keyboard Chinese input method and its key position arrangement | |
| CN1072785A (en) | Irrational rank-numeral synthetic coding method and keyboard thereof | |
| CN1145875C (en) | Chinese-character isomorphic input method for computer | |
| CN100339808C (en) | U Code Chinese character inputting method | |
| CN1570901A (en) | Hand-held interactive dictionary enquiry device and method | |
| CN1510554A (en) | Embedded applied Chinese character inputting method | |
| CN1202464C (en) | Single finger input system specific for handhold terminals | |
| CN1104673C (en) | Popularized Lin code inputting method for Chinese characters |
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 | ||
| C19 | Lapse of patent right due to non-payment of the annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |