CN103019397A - Multilingual-mixing text line input method and device - Google Patents
Multilingual-mixing text line input method and device Download PDFInfo
- Publication number
- CN103019397A CN103019397A CN2011102797946A CN201110279794A CN103019397A CN 103019397 A CN103019397 A CN 103019397A CN 2011102797946 A CN2011102797946 A CN 2011102797946A CN 201110279794 A CN201110279794 A CN 201110279794A CN 103019397 A CN103019397 A CN 103019397A
- Authority
- CN
- China
- Prior art keywords
- input
- character
- identification engine
- identification
- recognition result
- 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.)
- Granted
Links
Images
Abstract
The invention discloses a multilingual-mixing text line input method and device, and belongs to the field of input methods. The input method comprises the following steps of: invoking preset recognition engines of at least two types of languages to recognize the input chirography of a collected current character to obtain a recognition result; setting at least two character input areas; according to the obtained recognition result, setting recognition engines of different language categories respectively corresponding to the at least two character input areas; collecting the character input chirography; according to the input area in which the input chirography is positioned, invoking the recognition engine of the corresponding language to recognize the input chirography; and obtaining a recognition result. According to the method disclosed by the invention, the input areas corresponding to different recognition engines are given according to the recognition result of the character input by a user, the corresponding recognition engine can be quickly invoked to recognize the character chirography input into the corresponding input area to obtain the recognition result, and the process of manually and frequently switching recognition languages is omitted, thereby being convenient for users and improving the input efficiency.
Description
Technical field
The present invention relates to input method field, relate to a kind of multiple language input-system, particularly a kind of line of text input method and device of multilingual mixing.
Background technology
In existing most of input methods, the input of each language is all needed to provide the language input system of a correspondence.For example, the Chinese that existing input method all exists and English input when needs input Chinese, switch to the state of input in Chinese; When the needs input is English, namely switch to the state of english input.Therefore, when wherein a kind of of the multiple different language of input, need repeatedly between different input languages, carry out correct switching and use.
Not need the people be the multiple language input-system that carries out the switching of target language but also have, and is that the patent file of CN98109769 has just proposed one corresponding to multilingual input system such as application number.This multi-language system is font information and the format information that arranges according to application program, or identifies target language by the information that the environment in multiple language input-system arranges function setting.When the target language of input was Japanese, system used the Japanese processing section to carry out the Japanese input processing.When the target language of input was English, system made in English processing section execution English input processing.When target language was Chinese, system used Chinese processing partly to carry out the input processing of Chinese.
As seen, there is following defective at least in multilingual short sentence input of the prior art: for each line of text of input, system can only provide a kind of identification candidate of language, when needs are inputted the line of text of a multilingual mixing, need manually frequently to switch between the different language input language, the user uses inconvenience.
Summary of the invention
For the inconvenience that needs user's manual switchover input language to bring to the user in the line of text that solves the multilingual mixing of input, the invention provides a kind of line of text input method and device of multilingual mixing, can input in the process of line of text of multilingual mixing the user, the language classification of the text of automatic decision user input, thereby provide correct line of text recognition result, and do not need the user manually to switch the input language classification of text.
The line of text input method of a kind of multilingual mixing provided by the invention comprises: call the macaronic at least identification engine that presets the input person's handwriting of the current character of collection is identified, obtain recognition result; At least two character input areas are set, and according to the recognition result that obtains the respectively identification engine of corresponding different language classification of described at least two character input areas are set; Gather the input person's handwriting of character, according to the residing input area of described input person's handwriting, call the identification engine of corresponding language described input person's handwriting is identified, and obtain recognition result.
Further, described method also comprises: after definite line of text end of input, obtain the recognition result of the line of text of input according to the recognition result of the character script of each input and the language model set up in advance, and show.
On the other hand, the line of text input media of a kind of multilingual mixing provided by the invention comprises:
The initial identification module is used for calling the macaronic at least identification engine that presets the input person's handwriting of the single character of collection is identified, and obtains recognition result;
Input area arranges module, is used for arranging at least two character input areas, and according to the recognition result that obtains the respectively identification engine of corresponding different language classification of described at least two character input areas is set;
Judge identification module, be used for gathering the input person's handwriting of character, according to the residing input area of described input person's handwriting, call the identification engine of corresponding language described input person's handwriting is identified, and obtain recognition result.
Line of text input method and the device of a kind of multilingual mixing provided by the invention, recognition result according to the character of user input, adopt nearby principle, at least two input areas are set, the speech recognition engine that different input areas is corresponding different, the user inputs in corresponding input area, can call fast corresponding identification engine the character script of input is identified, and draw recognition result.Both guarantee the accuracy of the line of text identification of multilingual mixing, do not needed again the language classification of user's manual switchover input text in the process of the line of text of the multilingual mixing of input, convenient for users to use, improved input efficiency.
Description of drawings
Fig. 1 shows the multilingual mixing line of text input method process flow diagram of one embodiment of the present of invention;
Fig. 2 shows the multilingual mixing line of text input method inputting interface schematic diagram of one embodiment of the present of invention;
Fig. 3 shows the multilingual mixing line of text input method process flow diagram of an alternative embodiment of the invention;
Fig. 4 shows the schematic diagram of a character of inputting interface handwriting input shown in Figure 2;
Fig. 5 shows the input position prompting frame schematic diagram that inputting interface shown in Figure 4 provides;
Fig. 6 shows the schematic diagram of proximal most position place input in the input position prompting frame that Fig. 5 provides;
Fig. 7 shows the input position prompting frame schematic diagram that inputting interface shown in Figure 6 provides;
Fig. 8 shows the schematic diagram of highest distance position place input in the input position prompting frame that Fig. 7 provides;
Fig. 9 shows the schematic diagram that the inputting interface short sentence end of input that provides at Fig. 8 provides recognition result;
Figure 10 shows the multilingual mixing line of text input device structure figure of another embodiment of the present invention.
Embodiment
A kind of multilingual mixing line of text input method process flow diagram of the present invention as shown in Figure 1, comprising: call the macaronic at least identification engine that presets the input person's handwriting of the current character of collection is identified, obtain recognition result; At least two character input areas are set, and according to the recognition result that obtains the respectively identification engine of corresponding different language classification of described at least two character input areas are set; Gather the input person's handwriting of character, according to the residing input area of described input person's handwriting, call the identification engine of corresponding language described input person's handwriting is identified, and obtain recognition result.By method of the present invention, recognition result according to the character of user input, provide different input area corresponding to identification engine, the user inputs in corresponding input area, can call rapidly corresponding identification engine identifies the character of input, and draw recognition result, and made things convenient for the user, improved input efficiency.
The identification engine of the language that presets of the present invention comprises two kinds at least, as long as the character of several language that preset has obvious difference in the written handwriting feature, when carrying out multilingual line of text input, just can be with the identification engine of these several language as the speech recognition engine that presets, adopt method disclosed by the invention, realize the input of the line of text that these several language mix.The identification engine of the language that presets of the present invention is preferably two kinds, a kind of is that engine identified in Chinese characters, another kind is the western language character recognition engine, as: Chinese identification engine and English identification engine, Chinese identification engine and French identification engine, Korean identification engine and English identification engine, Japanese OCR engine and English identification engine etc.
Below by specific embodiment content of the present invention is described in further detail, but the present invention is not restricted to this.
In the electronic equipment with line of text input function (such as computer, mobile phone etc.), at first need to have display device, show inputting interface on the described display device, the inputting interface of embodiments of the invention comprises: person's handwriting input field 23, identification candidate regions 22 and text editing district 21.Wherein, person's handwriting input field 23 provides the zone of carrying out handwriting input to the user, and the user can be undertaken in this zone hand-written by energy contactor control devices such as mouse, pointer or fingers, the input character person's handwriting; Identification candidate regions 22 is that with the zone that the recognition result that obtains shows according to score order from high to low, the user can select the identification candidate in this zone after the identification engine was identified for the character script of user's input.This electronic equipment can send to this editing area 21 with the identification candidate of user selection; Text editing district 21 is positions that the target candidate of user selection shows.
Embodiment 1
In the present embodiment, to preset macaronic identification engine as example: a kind of is that engine identified in Chinese, another kind is English identification engine, describes the bilingual identification engine that presets in order to distinguish, and respectively this bilingual identification engine is called: the first identification engine, the second identification engine.In the present embodiment, Chinese identification engine is the first identification engine, and English identification engine is the second identification engine.Accordingly, language classification corresponding to the first identification engine is that first language, language classification corresponding to the second identification engine are second language.Be appreciated that also English can be identified engine is the first identification engine, Chinese identification engine is the second identification engine, and the present invention is not construed as limiting this.
A kind of multilingual mixing line of text input method that the present embodiment provides as shown in Figure 3, comprises following step.
At first, the character script information of Gather and input, and show.As shown in Figure 4, the user is in the person's handwriting input field of the inputting interface of electronic equipment 23 input illustrated contents (person's handwriting of " you "), and electronic equipment gathers the character script of user's input, and shows.After detecting the character script end of input, surpass default character input time such as detecting the mouse-up time, think the current character end of input.Electronic equipment calls the bilingual identification engine that presets the input person's handwriting of the character of collection is identified, and obtains recognition result.In the present embodiment, the bilingual identification engine that presets is respectively the first identification engine (Chinese identification engine) and the second identification engine (English identification engine), electronic equipment calls respectively described two identification engines that preset the character script information that gathers is identified, obtain two groups of recognition results, wherein, described recognition result comprises the score of identification candidate characters and this candidate characters, the identification engine flag, described identification candidate characters is according to from high to low arranged sequentially of score, score is higher, and the person's handwriting that input is described is that the confidence level of person's handwriting of this candidate characters is higher.As call the first identification engine and identify the recognition result that obtains and comprise: (" you ", 99.5,1), (" still ", 98.0,1), (" but ", 97.0,1) Calling the second identification engine identifies the recognition result that obtains and comprises: (" tip ", 0.15,2), (" Job ", 0.20,2), (" Jod ", 0.09,2) ...Wherein, " you " and " tip " etc. are the identification candidate characters; 99.5 and 0.15 be " you " and " tip " correspondence score; 1 represents the first identification engine, and 2 represent the second identification engine.
Because the score scope of the recognition result of two different language identification engines may be different, behind the recognition result that obtains respectively two different identification engines, two identification engine identifications need to be obtained identifying the Score Normalization of candidate characters, and the identification candidate characters that will obtain obtains the identification candidate character list according to from high to low arranged sequentially of score.And should identify candidate character list and temporarily preserve.Electronic equipment is Chinese or English according to the language classification of the candidate word that ranks the first in the definite identification candidate character list that obtains of the sign of corresponding identification engine, and the language classification of definite candidate word is the language classification of the character of current input.By judging the identification engine identification of character, identification engine corresponding to character that is easy to obtain identifying in the candidate character list is the first identification engine or the second identification engine (i.e. Chinese identification engine or English identification engine), and the present embodiment is not repeated herein.At least the score of the first two candidate characters has surpassed default score threshold value in the identification candidate character list, and identification engine corresponding to described at least the first two candidate characters is not simultaneously, determines that the identification engine of the default language classification that electronic equipment arranges is as the identification engine of the character of current input.Input content as shown in Figure 4, electronic equipment determine that identification engine corresponding to the character of current input is for being the first identification engine (i.e. Chinese identification engine).Described default score threshold value is definite according to the recognition training result of different engines.
In order to make things convenient for the user to input fast and to identify, the present invention is around the character of current input, next-door neighbour's current character, set gradually at least two character input areas, described at least two character input areas can be horizontally or the mapping relations of vertically arranging, and set in advance identification engine corresponding to described at least two character input areas and the corresponding identification engine of current input character.Preferably, the direction along the line of text input sets gradually at least two character input areas.When the line of text along continuous straight runs was inputted to the right, at least two character input areas of described setting set gradually to the right along the right side of the character of current input; When line of text was vertically inputted downwards, at least two character input areas of described setting set gradually downwards along the below of the character of current input.Also can work as the line of text along continuous straight runs when inputting to the right, at least two character input areas of described setting vertically arrange along the right side of the character of current input.
In the present embodiment, because the macaronic identification engine that presets is respectively Chinese identification engine (the first identification engine) and English identification engine (the second identification engine), be that the present embodiment is applicable to Chinese and the input of English mixing line of text, according to Chinese with English the line of text writing style---level is write from left to right, the character that is close to current input on the right side of the character of current input arranges two adjacent character input areas.The present embodiment is with the position of two character input areas of dashed rectangle indication, as shown in Figure 5.The present invention is not construed as limiting the quantity of character input area, and still, the quantity of character input area should be corresponding with the quantity of the identification engine that presets.The present invention also is not construed as limiting the form of expression of character input area, can be the dotted line frame, can also be underscore, bracket, different colours piece etc., writes the zone as long as can enough distinguish multilingual.The position relationship of two character input areas of described setting and the character of current input can Set arbitrarily, such as horizontal or vertical arrangement etc., can level arrange to the right, also can arrange straight down, can adjacent (as shown in this embodiment), also can be non-conterminous (as a character input area on the right side of the character of current input, character input area above the character of current input).Preferably, described two character input areas are horizontal on the adjacent right side of the character of current input.
According to the position of the character of current input and the size of input area, obtain and the position of record input area.Take interface shown in Figure 5 as example, the coordinate that 23 leftmost sides, person's handwriting input field are set is x=0, and rightmost side coordinate is x=240, and in the person's handwriting coordinate of the first character of input, the horizontal ordinate maximal value is x
1=40, the original width that the character input area is set is W
1=50, interval W between the horizontal character input area
2=5, the input area that then is close to an input character is that the horizontal ordinate scope of the first input area is (40,90), and the input area of an input character is that the horizontal ordinate scope of the second input area is (100,150) on the interval.For the convenience of input area is set, can there be the interval between the adjacent input area.
Identification engine corresponding to character based on the current input of having determined in step 32 is the first identification engine (i.e. Chinese identification engine), in this step, according to identification engine corresponding to the character of current input identification engine corresponding to described two character input areas is set.According to nearby principle, identification engine corresponding to character that identification engine corresponding to input area adjacent with the character of current input is current input is set; The setting speech recognition engine corresponding with the input area of the character pitch of current input is another identification engine.In the present embodiment, it is the first identification engine (i.e. Chinese identification engine) that speech recognition engine corresponding to input area adjacent with the character of current input is set, and calls the character script that the first identification engine is identified in input in this zone; The setting speech recognition engine corresponding with the input area of the character pitch of current input is the second identification engine (i.e. English identification engine), calls the second identification engine and is identified in the character script of inputting in this zone.When at least two character input areas that arrange vertically arrange along the right side of the character of current input, the corresponding relation of identification engine of the character of input area and current input can be set arbitrarily in advance, as it is identical to be arranged in identification engine corresponding to identification engine corresponding to the input area of top, character right side of current input and the character of current input.
Preferably, described variation according to written handwriting can corresponding size and the position of adjusting input area.During such as input character person's handwriting in the input area adjacent with the character of current input, the distance of the frame of position and the input area of last sampled point of the character script of input is during less than the threshold value preset, the adjustment away from the current character position can be made in the position of this frame, input area can be widened automatically as described, and the position of another input area adjacent with this zone can be moved to the direction away from current character accordingly.
At inputting interface as shown in Figure 5, the user can proceed input.As shown in Figure 6, when continuing input character in the input area of user's character of an input on the next-door neighbour, the character script of electronic equipment Gather and input, and according to the position of described character script and input area, determine the residing character input area of described character script.The concrete grammar of determining the residing character input area of described character script has a variety of, and for example: the input area that is in the starting point coordinate of character script is as the residing input area of the character of current input.As shown in Figure 6, the user is at electronic equipment inputing characters person's handwriting, and the horizontal ordinate of the starting point of the character script that collects when electronic equipment is 60, in the time of in the horizontal ordinate scope of the first input area, determines that the character of current input is in the first input area.Perhaps, determine the input area that the character of current input is in according to the coordinate figure scope of the person's handwriting point of this character and the coordinate figure scope degree of overlapping of input area.As shown in Figure 6, the degree of overlapping of the horizontal ordinate of the scope of the horizontal ordinate of the character of input and the first input area greater than with the degree of overlapping of the horizontal ordinate of the second input area, therefore, determine that the character of current input is in the first input area.The method of the residing input area of character of judging current input is a lot, enumerates no longer one by one at this.
After the character of determining current input is in the first input area, call character recognition engine corresponding to the first input area and will identify the character script of the current input of collection, obtain the recognition result of the character of current input: comprise identification candidate characters and score and the identification engine flag corresponding with each identification candidate characters.In this example, call Chinese identification engine corresponding to the first input area the character script of input is identified, the candidate characters that obtains is such as " very ".
According to the residing character input area of described character script, determine the identification engine corresponding to character of current input, electronic equipment is preserved the recognition result of the character of current input, and identification engine corresponding to the character of current input.In the present embodiment, be in the first input area because determine the character of current input in the step 34, therefore preserve identification engine corresponding to this input area, i.e. the first identification engine (Chinese identification engine) is the identification engine of the correspondence of the character of current input.
Repeated execution of steps 33 is to step 35, and the user can proceed the line of text input.The user in input " very " afterwards, electronic equipment execution in step 33, on " very " right side two adjacent input areas are set, and setting identification engine corresponding to first input area adjacent with the character " very " of current input is Chinese identification engine, the identification engine corresponding with second input area at character " very " interval of current input is English identification engine, and inputting interface as shown in Figure 7.Electronic equipment judges within default character input interval time whether still have the character script input, if do not have, then thinks the line of text end of input, execution in step 36; If have, then execution in step 34.
If the user need to input second language (namely English), then with the second input area input character person's handwriting at character " very " interval of current input, as shown in Figure 8, user's input be the person's handwriting of English word " nice ".In step 34, the character script of Gather and input according to the residing character input area of described character script, is called the i.e. English identification of corresponding identification engine engine the character script that gathers is identified, and obtains recognition result, such as " nice ".
When electronic equipment was judged within default character input interval time, when not having the character script input, then think the line of text end of input.Described default character input was set interval time as required, such as 1 second.Described default character input is greater than each interval time of input character interval time.After definite line of text end of input, electronic equipment obtains the recognition result of the line of text of input according to the language model of the recognition result of the character script of each input and in advance foundation, and shows.Wherein, the recognition result of the character script of each input obtains in step 31 and step 34.The detailed process of the recognition result of the line of text of acquisition input is as follows: each identification candidate of the first identification candidate of the first character of input and second character of input sets up respectively combination to be verified, inquires about the combined probability of the combination to be verified of setting up by the language model of setting up in advance again; And then the Selection and Constitute probability surpasses the combination that specifies number of default probability threshold value, sets up combination to be verified with the identification candidate of the 3rd character of input, the combined probability of the combination to be verified of setting up by the language model inquiry of setting up in advance again.By that analogy, obtain the candidate result of a plurality of combinations.According to the combined probability order from big to small of the candidate result of a plurality of combinations of the described acquisition recognition result in the line of text of the multilingual mixing of identification candidate regions 22 interior demonstrations inputs.Said " char " refers to a symbol corresponding to input person's handwriting herein, and its recognition result can be an English word or Chinese phrase, depends on the recognition result that corresponding identification engine provides.
Line of text input method by multilingual mixing provided by the invention, by method of the present invention, recognition result according to the character of user input adopts nearby principle, and at least two input areas are set, the speech recognition engine that different input areas is corresponding different, the user inputs in corresponding input area, can call fast corresponding identification engine the character of input is identified, and draw recognition result, make things convenient for the user, improved input efficiency.
Embodiment 2
On the other hand, the invention provides a kind of line of text input media of multilingual mixing, as shown in figure 10, this device comprises:
Input area arranges module 102, is used for arranging at least two character input areas, and according to the recognition result that obtains the respectively identification engine of corresponding different language classification of described at least two character input areas is set;
Wherein, described initial identification module 101 specifically is used for: the character script information of Gather and input, and show; After detecting the character script end of input, electronic equipment calls the bilingual identification engine that presets and respectively the input person's handwriting of the character that gathers is identified, and obtains recognition result.In the present embodiment, the bilingual identification engine that presets is respectively the first identification engine (Chinese identification engine) and the second identification engine (English identification engine), electronic equipment calls respectively described two identification engines that preset the character script information that gathers is identified, obtain two groups of recognition results, wherein, described recognition result comprises the score of identification candidate characters and this candidate characters, described identification candidate characters is according to from high to low arranged sequentially of score, score is higher, and the person's handwriting that input is described is that the confidence level of person's handwriting of this candidate characters is higher.
Described input area arranges module 102 and further comprises: position setting unit and identification engine setting unit.Wherein, described position setting unit is used for around the character of current input and along the direction of line of text input, the character that is close to current input sets gradually described two input areas at least.Be specially according to other writing style of class of languages that presets, at least two character input areas are set, the quantity of described input area is corresponding with the kind of the identification engine that presets, the form of expression of described input area, and the position relationship of the character of input area and current input is with the description of step 34 in the embodiment of the invention 1.Described identification engine setting unit is used for setting in advance the mapping relations of speech recognition engine corresponding to described at least two input areas identification engine corresponding with the character of current input; Comprise, language classification according to the character of current input arranges speech recognition engine corresponding to described two character input areas, according to nearby principle, identification engine corresponding to language classification that identification engine corresponding to input area adjacent with the character of current input is the character of current input is set; The setting speech recognition engine corresponding with the input area of the character pitch of current input is speech recognition engine corresponding to another language classification.In the present embodiment, it is the first identification engine (i.e. Chinese identification engine) that speech recognition engine corresponding to input area adjacent with the character of current input is set, and calls the character script that the first identification engine is identified in input in this zone; The setting speech recognition engine corresponding with the input area of the character pitch of current input is the second identification engine (i.e. English identification engine), calls the second identification engine and is identified in the character script of inputting in this zone.
Described judgement identification module 103 further comprises: person's handwriting collecting unit, region decision unit and recognition unit.Described person's handwriting collecting unit is used for gathering the character script of user's input.Described region decision unit is used for the position according to described character script and input area, determines the residing character input area of described character script.The concrete grammar of determining the residing character input area of described character script has a variety of, specifically can referring to the step 34 of the embodiment of the invention 1, enumerate no longer one by one at this.Described recognition unit is used for calling identification engine corresponding to definite input area will be identified the character script of the current input of collection, obtain the recognition result of the character of current input: comprise identification candidate characters and the score corresponding with each identification candidate characters.In this example, call Chinese identification engine corresponding to the first input area the character script of input is identified, the candidate characters that obtains is such as " very ".After obtaining recognition result, electronic equipment can preserve current input character recognition result and and the language classification of the character of current input, in order to continue to arrange input area.
Described input media also comprises and obtains display module 104, is used for when definite line of text end of input after, and the recognition result of the line of text of inputting according to the language model acquisition of the recognition result of the character script of each input and in advance foundation, and show.Concrete acquisition process is referring to the step 36 of the embodiment of the invention 1.
The device that provides by the embodiment of the invention, recognition result according to the character of user input, provide different input area corresponding to identification engine, the user inputs in corresponding input area, can call rapidly corresponding identification engine identifies the character of input, and draw recognition result, and made things convenient for the user, improved input efficiency.
Embodiments of the invention also disclose a kind of electronic equipment, comprise contactor control device, display device, also comprise the line of text input media of embodiment 2 described a kind of multilingual mixing.
Although invention has been described in conjunction with the embodiments for the above, to those skilled in the art, be understandable that, above-described embodiment should not thought the protection domain that limits the present invention with this only for example purposes.Protection scope of the present invention is only limited by claim.Although in the above-mentioned explanation, for ease of understanding, the step of method has been adopted the succession description, should be pointed out that for the order of above-mentioned steps and do not do strict restriction.In addition, realize that all or part of step in above-described embodiment method is to come the relevant hardware of instruction to finish by program, this program can be stored in the computer read/write memory medium, as: ROM/RAM, magnetic disc, CD etc.Also have, the apparatus structure shown in accompanying drawing or the embodiment only is to schematically show logical organization.The module that wherein shows as separating component may or may not be physically to separate, and the parts that show as module also may be or may not be physical modules.
In addition, in claims, term " comprises " does not get rid of other parts of existence or step.In addition, although comprised each feature in different claims, these features can be advantageously combined, and each feature is included in the different claims and does not mean that the combination of these features is infeasible and/or disadvantageous.In addition, single implication is not got rid of a plurality of.Therefore, the implication of " " etc. is not got rid of a plurality of.
In addition, those skilled in the art can make amendment, change or replace it under the prerequisite that does not break away from spirit of the present invention, and still, the various equivalent variations according to the present invention does still belong to the scope that the present invention is contained.
Claims (11)
1. the line of text input method of a multilingual mixing is characterized in that, comprising:
Call the macaronic at least identification engine that presets the input person's handwriting of the current character of collection is identified, obtain recognition result;
At least two character input areas are set, and according to the recognition result that obtains the respectively identification engine of corresponding different language classification of described at least two character input areas are set;
Gather the input person's handwriting of character, according to the residing input area of described input person's handwriting, call the identification engine of corresponding language described input person's handwriting is identified, and obtain recognition result;
Wherein, described recognition result comprises the identification candidate characters, and corresponding score and the identification engine flag of described identification candidate characters.
2. input method according to claim 1, it is characterized in that, also comprise: after definite line of text end of input, obtain the recognition result of the line of text of input according to the recognition result of the character script of each input and the language model set up in advance, and show.
3. input method according to claim 1, it is characterized in that, described at least two character input areas are around the character of current input, the character that is close to current input sets gradually, and sets in advance the mapping relations of identification engine corresponding to described at least two the character input areas identification engine corresponding with the character of current input.
4. input method according to claim 1 is characterized in that, the described macaronic at least identification engine that presets comprises: Chinese characters identification engine and western language identification engine, and the character input area of setting is two,
Wherein, described Chinese characters identification engine comprises: Chinese identification engine, Japanese OCR engine, Korean identification engine; Described western language character recognition engine comprises: English identification engine.
5. according to claim 1 or 3 described input methods, it is characterized in that, described character input area is two, and on the right side of the character of current input, the character level that is close to current input sets gradually to the right.
6. input method according to claim 1 is characterized in that, described recognition result according to obtaining arrange described at least two character input areas respectively the identification engine of corresponding different language classification be specially:
The identification candidate characters that obtains according to from high to low arranged sequentially of score, is obtained the identification candidate character list;
According to the identification engine flag of candidate characters, identification engine corresponding to candidate characters that ranks the first in the identification candidate character list of determining to obtain, and determine identification engine corresponding to character that this identification engine is current input;
Identification engine corresponding to character that identification engine corresponding to input area adjacent with the character of current input is current input is set; The setting speech recognition engine corresponding with the input area of the character pitch of current input is another identification engine.
7. input method according to claim 1 is characterized in that, the form of expression of described character input area comprises: dotted line frame, underscore, bracket, different colours piece.
8. the line of text input media of a multilingual mixing is characterized in that, comprising:
The initial identification module is used for calling the macaronic at least identification engine that presets the input person's handwriting of the single character of collection is identified, and obtains recognition result;
Input area arranges module, is used for arranging at least two character input areas, and according to the recognition result that obtains the respectively identification engine of corresponding different language classification of described at least two character input areas is set;
Judge identification module, be used for gathering the input person's handwriting of character, according to the residing input area of described input person's handwriting, call the identification engine of corresponding language described input person's handwriting is identified, and obtain recognition result.
9. input media according to claim 8 is characterized in that, described input area arranges module and further comprises: position setting unit and identification engine setting unit,
Wherein, the setting unit character that is used for being close to current input around the character of current input in described position sets gradually described at least two character input areas;
Described identification engine setting unit is used for setting in advance the mapping relations of speech recognition engine corresponding to described at least two the character input areas identification engine corresponding with the character of current input; Comprise, according to identification engine corresponding to the character of current input speech recognition engine corresponding to described two character input areas is set, comprises: identification engine corresponding to language classification that identification engine corresponding to input area adjacent with the character of current input is the character of current input is set.
10. input media according to claim 8, it is characterized in that, also comprise: obtain display module, the described display module that obtains is used for after definite line of text end of input, obtain the recognition result of the line of text of input according to the recognition result of the character script of each input and the language model set up in advance, and show.
11. an electronic equipment comprises contactor control device, display device, it is characterized in that, also comprises the line of text input media such as each described a kind of multilingual mixing of claim 8 to 10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110279794.6A CN103019397B (en) | 2011-09-20 | 2011-09-20 | A kind of line of text input method of multilingual mixing and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110279794.6A CN103019397B (en) | 2011-09-20 | 2011-09-20 | A kind of line of text input method of multilingual mixing and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103019397A true CN103019397A (en) | 2013-04-03 |
| CN103019397B CN103019397B (en) | 2015-11-25 |
Family
ID=47968080
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110279794.6A Active CN103019397B (en) | 2011-09-20 | 2011-09-20 | A kind of line of text input method of multilingual mixing and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103019397B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103488488A (en) * | 2013-09-26 | 2014-01-01 | 贝壳网际（北京）安全技术有限公司 | Text input check method, device ad mobile terminal |
| CN104951100A (en) * | 2015-06-23 | 2015-09-30 | 小米科技有限责任公司 | Character input method and device |
| TWI511030B (en) * | 2014-04-15 | 2015-12-01 | Acer Inc | Method for user interface display and electronic device using the same |
| CN106940596A (en) * | 2016-01-04 | 2017-07-11 | 北京峰盛博远科技股份有限公司 | A kind of recognition methods of multiple characters of handwriting input and system |
| CN110858317A (en) * | 2018-08-24 | 2020-03-03 | 北京搜狗科技发展有限公司 | Handwriting recognition method and device |
| CN117251061A (en) * | 2023-11-17 | 2023-12-19 | 广东省电信规划设计院有限公司 | AIGC-based on-screen intelligent input method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1203398A (en) * | 1997-06-11 | 1998-12-30 | 株式会社东芝 | Multiple language input-system |
| CN101515211A (en) * | 2008-02-22 | 2009-08-26 | 汉王科技股份有限公司 | Handwriting input method, device and user terminal |
| CN101882025A (en) * | 2010-06-29 | 2010-11-10 | 汉王科技股份有限公司 | Hand input method and system |
| CN102103434A (en) * | 2011-03-09 | 2011-06-22 | 王岩泽 | Handwriting input device and method for characters |
| WO2011102689A2 (en) * | 2010-02-19 | 2011-08-25 | Soon Jo Woo | Multilingual key input apparatus and method thereof |
-
2011
- 2011-09-20 CN CN201110279794.6A patent/CN103019397B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1203398A (en) * | 1997-06-11 | 1998-12-30 | 株式会社东芝 | Multiple language input-system |
| CN101515211A (en) * | 2008-02-22 | 2009-08-26 | 汉王科技股份有限公司 | Handwriting input method, device and user terminal |
| WO2011102689A2 (en) * | 2010-02-19 | 2011-08-25 | Soon Jo Woo | Multilingual key input apparatus and method thereof |
| CN101882025A (en) * | 2010-06-29 | 2010-11-10 | 汉王科技股份有限公司 | Hand input method and system |
| CN102103434A (en) * | 2011-03-09 | 2011-06-22 | 王岩泽 | Handwriting input device and method for characters |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103488488A (en) * | 2013-09-26 | 2014-01-01 | 贝壳网际（北京）安全技术有限公司 | Text input check method, device ad mobile terminal |
| TWI511030B (en) * | 2014-04-15 | 2015-12-01 | Acer Inc | Method for user interface display and electronic device using the same |
| CN104951100A (en) * | 2015-06-23 | 2015-09-30 | 小米科技有限责任公司 | Character input method and device |
| CN106940596A (en) * | 2016-01-04 | 2017-07-11 | 北京峰盛博远科技股份有限公司 | A kind of recognition methods of multiple characters of handwriting input and system |
| CN110858317A (en) * | 2018-08-24 | 2020-03-03 | 北京搜狗科技发展有限公司 | Handwriting recognition method and device |
| CN117251061A (en) * | 2023-11-17 | 2023-12-19 | 广东省电信规划设计院有限公司 | AIGC-based on-screen intelligent input method and device |
| CN117251061B (en) * | 2023-11-17 | 2024-03-15 | 广东省电信规划设计院有限公司 | AIGC-based on-screen intelligent input method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103019397B (en) | 2015-11-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103019397B (en) | A kind of line of text input method of multilingual mixing and device | |
| US8363949B2 (en) | Character recognition for overlapping textual user input | |
| CN103294363B (en) | A kind of searching method and terminal | |
| EP2650766B1 (en) | Multi-character continuous handwriting input method | |
| CN100587660C (en) | Method and device for forecasting and discriminating hand-written characters | |
| CN101576783B (en) | User interface, equipment and method for hand input | |
| US20140143721A1 (en) | Information processing device, information processing method, and computer program product | |
| CN102135838A (en) | Method and system for partitioned input of handwritten character string | |
| WO2014099948A1 (en) | Composition of handwritten messages on mobile computing devices | |
| CA2853199A1 (en) | Extracting principal content from web pages | |
| EP2897055A1 (en) | Information processing device, information processing method, and program | |
| CN102063620A (en) | Handwriting identification method, system and terminal | |
| CN102073884A (en) | Handwriting recognition method, system and handwriting recognition terminal | |
| CN103092343A (en) | Control method based on camera and mobile terminal | |
| KR101927064B1 (en) | Apparus and method for generating summary data about e-book | |
| CN102301366A (en) | Information processing device | |
| CN102243708B (en) | Handwriting recognition method, handwriting recognition system and handwriting recognition terminal | |
| CN102221976A (en) | Method for quickly inputting words based on incomplete identification | |
| CN102981767B (en) | A kind of individual Chinese character hand-written recognition method and system | |
| CN102981693A (en) | Multi-language hand writing input method and device | |
| CN102193718A (en) | Chinese character handwriting input method and Chinese character handwriting input system | |
| CN101882025A (en) | Hand input method and system | |
| CN108062301A (en) | Character translation method and its device | |
| CN101452368B (en) | Hand-written character input method | |
| CN104731766A (en) | Alphabetic writing lexicon establishing method, alphabetic writing lexicon establishing device, inputting method and inputting system |
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 |