CN105549890A - One-dimensional handwritten character input equipment and one-dimensional handwritten character input equipment - Google Patents

One-dimensional handwritten character input equipment and one-dimensional handwritten character input equipment Download PDF

Info

Publication number
CN105549890A
CN105549890A CN201511009156.7A CN201511009156A CN105549890A CN 105549890 A CN105549890 A CN 105549890A CN 201511009156 A CN201511009156 A CN 201511009156A CN 105549890 A CN105549890 A CN 105549890A
Authority
CN
China
Prior art keywords
character
dimension
gesture
stroke
user
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
Application number
CN201511009156.7A
Other languages
Chinese (zh)
Other versions
CN105549890B (en
Inventor
喻纯
孙科
钟鸣远
李心成
史元春
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tsinghua University
Original Assignee
Tsinghua University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tsinghua University filed Critical Tsinghua University
Priority to CN201511009156.7A priority Critical patent/CN105549890B/en
Publication of CN105549890A publication Critical patent/CN105549890A/en
Priority to PCT/CN2016/105694 priority patent/WO2017114002A1/en
Application granted granted Critical
Publication of CN105549890B publication Critical patent/CN105549890B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
    • G06F3/0487—Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser
    • G06F3/0488—Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser using a touch-screen or digitiser, e.g. input of commands through traced gestures
    • G06F3/04883—Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser using a touch-screen or digitiser, e.g. input of commands through traced gestures for inputting data by handwriting, e.g. gesture or text
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/02—Input arrangements using manually operated switches, e.g. using keyboards or dials
    • G06F3/023—Arrangements for converting discrete items of information into a coded form, e.g. arrangements for interpreting keyboard generated codes as alphanumeric codes, operand codes or instruction codes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/02—Input arrangements using manually operated switches, e.g. using keyboards or dials
    • G06F3/023—Arrangements for converting discrete items of information into a coded form, e.g. arrangements for interpreting keyboard generated codes as alphanumeric codes, operand codes or instruction codes
    • G06F3/0233—Character input methods
    • G06F3/0236—Character input methods using selection techniques to select from displayed items
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
    • G06F3/0487—Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser
    • G06F3/0488—Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser using a touch-screen or digitiser, e.g. input of commands through traced gestures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/10—Character recognition
    • G06V30/32—Digital ink

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Human Computer Interaction (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Multimedia (AREA)
  • Character Discrimination (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

The invention provides a one-dimensional handwritten character input method and equipment. The one-dimensional handwritten character input equipment comprises a user input interface, a detection unit, a character template database and a recognition unit, wherein the user input interface is provided for users to making a one-dimensional character gesture, wherein the one-dimensional character gesture represents reciprocating rowing on a straight line; the detection unit is configured to detect the one-dimensional character gesture and convert the one-dimensional character gesture into a one-dimensional signal; the character template database is configured to store the template of each character, wherein the templates of the characters corresponds to one-dimensional strokes and form a set of one or more sub-strokes according to a predetermined sequence, and the sub-strokes are basically line segments on the straight line; and the recognition unit is configured to receive the one-dimensional signal from the detection unit, convert the one-dimensional signal into a to-be-processed one-dimensional stroke and recognize the one-dimensional stroke into a corresponding character on the basis of the to-be-processed one-dimensional stroke and the template of each character stored in the character template database. The invention provides a character input method and equipment particularly suitable for carrying out input in relatively long and narrow one-dimensional input spaces.

Description

One dimension handwriting input equipment and one dimension hand-written character input method
Technical field
The present invention relates generally to handwritten text input technology, relates more specifically to one dimension handwriting input equipment and method.
Background technology
Along with developing rapidly of smart machine, the demand that smart machine carries out Text Input is also strong all the more.But smart machine is often due to the restriction of its physical aspect size, thus inputting interface is limited, make the Text Input mode of physically based deformation keyboard or soft keyboard no longer applicable, the Text Input on these smart machines becomes difficulty.The example of these smart machines such as has intelligent watch, intelligent glasses and Intelligent bracelet etc.
Text entry method can be divided into haply based on the method selected and the method based on handwritten form input.Such as provide physics or virtual keyboard based on the method selected, each key represents each letter or number, selecting (such as by clicking corresponding keys) certain letter, namely inputting this letter, selecting multiple letter to form word.Based on the method for handwritten form input, then the general form being substituted body by method simulation character (or word) of drawing or the corresponding character (or word) defined, is carried out identifying the text obtaining inputting by intellectual computing device.
Propose the technology of some input texts on limited area interface, the method to alphabetic coding of such as document 1; The one stroke gesture of document 2,3; The fuzzy keyboard of document 4; The gesture keyboard etc. of document 5,6.But most of technology be all for two dimension decoding or there is the interface of limited button.
Document 1:MacKenzie, I., Soukoreff, R.andHelga, J.1thumb, 4buttons, 20wordsperminute:DesignandevaluationofH4-Writer.UIST'11, (2011), 471-480.
PenComputingMagazine, pp.30-31.
Document 3:Gu, Z., Xu, X., Chu, C.andZhang, Y.ToWritenotSelect, aNewTextEntryMethodUsingJoystick.Human-ComputerInteracti on:InteractionTechnologies, (2015), 35-43.
Document 4:Poirier, F.andBelatar, M.UniGlyph:onlyonekeystrokepercharacterona4-buttonminima lkeypadforkey-basedtextentry.HCIInternational2007, (2007), 479-483.
Document 5:Kristensson, P.andZhai, S.SHARK2:alargevocabularyshorthandwritingsystemforpen-ba sedcomputers.UIST'04, (2004), 43-52.
Document 6:Zhai, S.andKristensson, P.Shorthandwritingonstyluskeyboard.CHI'03, (2003), 97-104.
Summary of the invention
The present invention pays close attention to the method based on handwritten form input, especially focuses on the method based on handwritten form input on one dimension interface.
The present invention desirable to provide limited inputting interface---such as on one dimension inputting interface---carry out handwriting input and identify method and apparatus.The example of described one dimension inputting interface such as has: the leg of spectacles surface along its length of intelligent glasses, the side frame of smart mobile phone, the side etc. of Intelligent bracelet.
According to an aspect of the present invention, provide a kind of one dimension handwriting input equipment, can comprise: user's inputting interface, manually make one dimension character gesture in chronological order in the mode contacting this user's inputting interface for user with body part or instrument of drawing, so-called one dimension character gesture is reciprocal paddling point-blank substantially; Detecting unit, be configured to the one dimension character gesture that detection user makes on user's inputting interface, transfer this one dimension character gesture to one-dimensional signal, described one-dimensional signal is the signal only having value in a coordinate system in a dimension; Character mother plate database, is configured to the template storing each character, and the template of each character corresponds to one dimension stroke, is the set of the one or more sub-stroke according to predefined procedure, and each sub-stroke is line segment substantially point-blank; Recognition unit, configuration receives this one-dimensional signal from detecting unit, this one-dimensional signal is converted to pending one dimension stroke, and this one dimension stroke recognition is respective symbols by the template of each character stored in the one dimension stroke pending based on this and Character mother plate database.
One dimension handwriting input equipment according to the embodiment of the present invention can also comprise: display unit; Wherein, at least one Character mother plate stored in Character mother plate database corresponds to multiple character, wherein, when processor identification one dimension stroke corresponds to multiple character as character candidates, processor performs process and makes to show the plurality of character candidates over the display, and wherein one of multiple character candidates is highlighted at any one time; And
Detect in response to detecting unit the selection gesture that user makes, the current character candidates be highlighted of processor selection is as character identification result; And detecting in response to detecting unit the mobile gesture that user makes, processor switches the character candidates be highlighted.
According to the one dimension handwriting input equipment of the embodiment of the present invention, wherein detect that gesture separated in the character that user makes in response to detecting unit, continuous print one dimension character gesture can distinguish by processor, to terminate the input of last character, and prepares the input receiving next character.
According to the one dimension handwriting input equipment of the embodiment of the present invention, gesture separated in wherein said character can be pause gesture, and pause gesture corresponds to user's finger or instrument of drawing dead time on user's inputting interface exceedes predetermined threshold.
According to the one dimension handwriting input equipment of the embodiment of the present invention, so-called one dimension character gesture can correspond to single stroke, single stroke is single continuous print stroke, makes in process in one dimension character gesture, and user's finger or instrument of drawing do not leave user's inputting interface.
According to the one dimension handwriting input equipment of the embodiment of the present invention, the sub-stroke of the stroke of the template of each character can have length and direction, and wherein length is one of two kinds of predetermined lengths.
According to the one dimension handwriting input equipment of the embodiment of the present invention, described one dimension character gesture can with reality two dimension character hand-written tendency similar process.
According to the one dimension handwriting input equipment of the embodiment of the present invention, at least one in the Character mother plate stored in Character mother plate database can correspond to multiple character.
According to the one dimension handwriting input equipment of the embodiment of the present invention, for the Character mother plate of specific character, can first by this character 90-degree rotation, then by hand-written for two dimension hand modelling to one dimension character gesture.
According to the one dimension handwriting input equipment of the embodiment of the present invention, this one-dimensional signal is converted to pending one dimension stroke and can comprises by described processor: identify flex point, flex point refers to carry out in time in this one dimension stroke forming process the point that stroke direct of travel changes, and flex point comprises starting point and the end point of stroke; Remove the flex point being less than intended pixel length the flex point middle distance time in front flex point; Line segment is divided based on remaining flex point; Get rid of the line segment that length in line segment is less than predetermined threshold, residue line segment is normalized, obtains the sequence of set as the sub-stroke according to time sequencing of the line segment after normalization.
According to the one dimension handwriting input equipment of the embodiment of the present invention, the template of described each character based on storing in this pending one dimension stroke and Character mother plate database, by this one dimension stroke recognition for respective symbols can comprise: the number determining the sub-stroke that this one dimension stroke comprises; Searching character template database, determines that template neutron number of strokes equals the template of the number of the sub-stroke that described this one dimension stroke determined comprises;
Each in determined template, according to formation order, determines that every sub-stroke is rendered as the probability of the sub-stroke of correspondence in described one dimension stroke, calculates the probability that this template given is rendered as described one dimension stroke; And the probability of described one dimension stroke is rendered as based on each template, be corresponding character by described one dimension stroke recognition.
According to the one dimension handwriting input equipment of the embodiment of the present invention, can also display unit being comprised, being configured to when there is multiple character candidates, with a visual form display predetermined number character candidates; In response to detecting that user makes one dimension character gesture and terminates to determine gesture, show the determined character candidates corresponding with this one dimension character gesture, described one dimension character gesture terminates to determine that gesture instruction one dimension character gesture terminates, and in response to detecting that user makes character and selects gesture, select the current candidate characters be highlighted, and be presented at the viewing area of the text of indicative input, or in response to detecting that user makes candidate characters and switches gesture, switching the current candidate characters that is highlighted and being switched to highlighted candidate characters with highlighted fashion display.
According to the one dimension handwriting input equipment of the embodiment of the present invention, wherein said candidate characters switches gesture can for finger be along from the current candidate characters be highlighted to the movement on the direction of the candidate characters that will select, in one dimension character gesture end point close to user's inputting interface on one-dimensional square the first border upwards, candidate characters is shown along from the first border to the direction of the second boundary according to probability order from high to low.
According to the one dimension handwriting input equipment of the embodiment of the present invention, described one dimension character gesture terminates to determine that gesture can finger is motionless on user's inputting interface reaches the schedule time in order to keep, described character selects gesture for be lifted from user's inputting interface by finger, and it is that moveable finger reaches preset distance that described candidate characters switches gesture.
According to the one dimension handwriting input equipment of the embodiment of the present invention, can also comprise: language model data storehouse, store the probability distribution of instruction by character, assign probability to the information of the word be made up of character; Described processor is under Text region state, switch between character input mode and word input pattern, wherein under word input pattern, reference language model database calculates the probability inputting concrete vocabulary under given one dimension stroke sequence condition in real time, based on the probability of the concrete vocabulary of input, candidate's vocabulary is sorted, and display on the display unit, selects for user.
According to the one dimension handwriting input equipment of the embodiment of the present invention, under vocabulary pattern, the graphic user interface of display unit comprises three regions: text filed, for showing the text of input; Character zone, for the character of Identification display; And word region, for the word candidate of Identification display.
According to the one dimension handwriting input equipment of the embodiment of the present invention, wherein said text filed height can be mapped to the one-dimensional square whole length upwards of user's inputting interface in one dimension character gesture, described text filed horizontal direction represents time shaft, and in described text filed rise time axle display one dimension character gesture, thus provide the two dimensional visible feedback of one dimension character gesture.
According to the one dimension handwriting input equipment of the embodiment of the present invention, described one dimension handwriting input equipment can be intelligent wearable device.
According to a further aspect in the invention, provide the one dimension hand-written character input method that a kind of one dimension handwriting input equipment performs, one dimension handwriting input equipment can comprise user's inputting interface and Character mother plate database, described one dimension hand-written character input method comprises: manually make one dimension character gesture in chronological order in the mode contacting user's inputting interface in response to user with finger or instrument of drawing, detect the one dimension character gesture that user makes on user's inputting interface, transfer this one dimension character gesture to one-dimensional signal, described one-dimensional signal is the signal only having value in a coordinate system in a dimension, so-called one dimension character gesture is reciprocal paddling point-blank substantially, receive this one-dimensional signal from detecting unit, this one-dimensional signal is converted to pending one dimension stroke, the template of each character stored in the one dimension stroke pending based on this and Character mother plate database, be respective symbols by this one dimension stroke recognition, wherein the template of each character corresponds to one dimension stroke, for the set of the one or more sub-stroke according to predefined procedure, each sub-stroke is line segment substantially point-blank.
According to the one dimension hand-written character input method of the embodiment of the present invention, wherein, at least one Character mother plate stored in Character mother plate database corresponds to multiple character, wherein, when recognizing one dimension stroke and corresponding to multiple character as character candidates, perform process to make to show the plurality of character candidates over the display, wherein one of multiple character candidates is highlighted at any one time; And detect in response to detecting unit the selection gesture that user makes, select the current character candidates be highlighted as character identification result; And detecting in response to detecting unit the mobile gesture that user makes, processor switches the character candidates be highlighted.
According to the one dimension hand-written character input method of the embodiment of the present invention, wherein detect that gesture separated in the character that user makes in response to detecting unit, continuous print one dimension character gesture is distinguished, to terminate the input of last character, and prepares the input receiving next character.
According to the one dimension hand-written character input method of the embodiment of the present invention, so-called one dimension character gesture can correspond to single stroke, single stroke is single continuous print stroke, makes in process in one dimension character gesture, and user's finger or instrument of drawing do not leave user's inputting interface.
According to the one dimension hand-written character input method of the embodiment of the present invention, the sub-stroke of the stroke of the template of each character can have length and direction, and wherein length can be one of two kinds of predetermined lengths.
According to the one dimension hand-written character input method of the embodiment of the present invention, described one dimension character gesture can with reality two dimension character hand-written tendency similar process.
According to the one dimension hand-written character input method of the embodiment of the present invention, describedly this one-dimensional signal is converted to pending one dimension stroke and can comprises: identify flex point, refer to carry out in time in this one dimension stroke forming process the point that stroke direct of travel changes, flex point comprises starting point and the end point of stroke; Remove the flex point being less than intended pixel length the flex point middle distance time in front flex point; Line segment is divided based on remaining flex point; Get rid of length in head and the tail line segment and be less than the line segment of predetermined threshold, residue line segment is normalized, obtains the sequence of set as the sub-stroke according to time sequencing of the line segment after normalization.
According to the one dimension hand-written character input method of the embodiment of the present invention, the template of described each character based on storing in this pending one dimension stroke and Character mother plate database, by this one dimension stroke recognition for respective symbols can comprise: the number determining the sub-stroke that this one dimension stroke comprises; Searching character template database, determines that template neutron number of strokes equals the template of the number of the sub-stroke that described this one dimension stroke determined comprises; Each in determined template, according to formation order, determines that every sub-stroke in this template is rendered as the probability of the sub-stroke of correspondence in described one dimension stroke, calculates the probability that described one dimension stroke corresponds to this template; And the probability of each template is corresponded to based on described one dimension stroke, be corresponding character by described one dimension stroke recognition.
According to the one dimension hand-written character input method of the embodiment of the present invention, can also comprise: language model data storehouse, store the probability distribution of instruction by character, assign probability to the information of the word be made up of character; Described processor is under Text region state, switch between character input mode and word input pattern, wherein under word input pattern, reference language model database calculates the probability inputting specific vocabulary under given one dimension stroke sequence condition in real time, based on the probability of the specific vocabulary of input, candidate's vocabulary is sorted, and display on the display unit, selects for user.
According to the one dimension hand-written character input method of the embodiment of the present invention, can under vocabulary pattern, on the graphic user interface of display unit: the text showing input on text filed; The character of Identification display on character zone; And on word region the word candidate of Identification display.
According to the one dimension hand-written character input method of the embodiment of the present invention, wherein said text filed height mapping is to the one-dimensional square whole length upwards of user's inputting interface in one dimension character gesture, described text filed horizontal direction represents time shaft, described hand-written recognition method can also comprise: in described text filed rise time axle display one dimension character gesture, thus provide the two dimensional visible of one dimension character gesture to feed back.
According to a further aspect in the invention, provide a kind of one dimension handwriting input equipment, can comprise: user's inputting interface, manually make one dimension character gesture in chronological order in the mode contacting this user's inputting interface for user with body part or instrument; Detecting unit, be configured to the one dimension character gesture that detection user makes on user's inputting interface, transfer this one dimension character gesture to one-dimensional signal, described one-dimensional signal is the signal only having value in a coordinate system in a dimension; Character mother plate database, is configured to the template storing each character, and the template of each character is the proper vector of described one-dimensional signal being carried out to statistically feature extraction acquisition; And processor, configuration receives from this one-dimensional signal of detecting unit, and this one-dimensional signal is converted to pending proper vector, and the template of each character stored in the proper vector pending based on this and Character mother plate database, identifies and obtain respective symbols.
According to the one dimension handwriting input equipment of the embodiment of the present invention, described one dimension character gesture can be one of following items: the reciprocal paddling substantially point-blank carried out on described user's inputting interface; The single-point that described user's inputting interface carries out presses down; And the rotation on the one dimension angle coordinate system carried out on described user's inputting interface.
According to one dimension hand-written character input method and the equipment of the embodiment of the present invention, be particularly suitable for carrying out text event detection in the relatively long and narrow one dimension input space.
Accompanying drawing explanation
Below in conjunction with accompanying drawing in the detailed description of the embodiment of the present invention, these and/or other side of the present invention and advantage will become clearly and be easier to understand, wherein:
Fig. 1 shows the simplification structured flowchart of the one dimension handwriting input equipment 100 according to the embodiment of the present invention.
Fig. 2 shows the sub-stroke pie graph of each character of Character mother plate according to an embodiment of the invention.
Fig. 3 shows the distribution of the sub-stroke of short, medium, and long in the actual gesture made of user after normalization.
The sub-stroke that Fig. 4 shows each character of Character mother plate according to another embodiment of the present invention forms schematic diagram.
Fig. 5 shows the process flow diagram of the one dimension hand-written character input method performed by one dimension handwriting input equipment according to the embodiment of the present invention.
Fig. 6 shows the example of the graphic user interface according to the embodiment of the present invention.
Fig. 7 shows process flow diagram one-dimensional signal corresponding for one dimension character gesture being converted to the method 400 of pending one dimension stroke according to the embodiment of the present invention.
This one dimension stroke recognition is the process flow diagram of the method 500 of respective symbols by Fig. 8 template shown based on each character stored in pending one dimension stroke and Character mother plate database.
Fig. 9 shows under words enter pattern, utilizes the process flow diagram of the illustrative methods 600 of the vocabulary of bayes method and language model prediction input.
Figure 10 (a), (b), (c) show the example that the one dimension handwriting input equipment that can be used as Fig. 1 realizes the real world devices of environment.
Embodiment
In order to make those skilled in the art understand the present invention better, below in conjunction with the drawings and specific embodiments, the present invention is described in further detail.
The following describes the implication of term used herein.
" one dimension character gesture ", reciprocal paddling point-blank, the length only in this rectilinear direction and directional information, the not degree of depth vertical with this rectilinear direction and elevation information.Namely the gesture that toilet is made is not strict point-blank carrying out, and only extracts length in the straight direction and directional information yet, and ignores the information in the degree of depth and height.
" one-dimensional signal " refers to the signal only having value in a coordinate system in a dimension.
" one dimension stroke " is the set of the one or more sub-stroke according to predefined procedure, and each sub-stroke is line segment substantially point-blank.
" single stroke ": single continuous print stroke, makes in process in one dimension character gesture, and user's finger or instrument of drawing do not leave user's inputting interface.
One, one dimension handwriting input equipment
Fig. 1 shows the simplification structured flowchart of the one dimension handwriting input equipment 100 according to the embodiment of the present invention.
As shown in Figure 1, one dimension handwriting input equipment 100 can comprise: user's inputting interface 110, detecting unit 120, Character mother plate database 130, recognition unit 140.
Figure 10 gives the example that the one dimension handwriting input equipment that can be used as Fig. 1 realizes the real world devices of environment, and wherein Figure 10 (a) is intelligent glasses, Figure 10 (b) is smart mobile phone, Figure 10 (c) is intelligent watch.
Get back to Fig. 1, user's inputting interface 110 manually makes one dimension character gesture in the mode contacting this user's inputting interface for user in chronological order with finger or instrument of drawing, and so-called one dimension character gesture is reciprocal paddling point-blank substantially.
Here the example of user's inputting interface 110 can be the side of the temple of intelligent glasses, the side frame of smart mobile phone, the dial plate frame etc. of intelligent watch.User's inputting interface 110 can be approximated to be straight line.But as expansion, although such as bracelet on the surface along the slip of annular be not straight line, but this annular can expand into straight line, the one dimension character gesture of the embodiment of the present invention also can be carried out on such loop wire, and therefore such ring user interface is also within covering scope of the present invention.
It should be noted that, " basic is reciprocal paddling point-blank " here even if represent that one dimension character gesture is not strictly point-blank, but as long as in fact point-blank, is also applicable to the present invention.Such as, as the first sub-stroke at straight line A along a direction paddling, second sub-stroke is back paddling from the first sub-stroke end, even if the line segment corresponding to the second sub-stroke is not strictly be positioned on straight line A, also the present invention's " one dimension character gesture " mentioned here is belonged to, also be only extracted in the signal (signal along straight line A) in a dimension, and be not extracted the signal of another dimension (signal of such as vertical with straight line A axle).This meets user's actual act situation, and namely the present invention allows user to point has slight departing from single coordinate axis.
Here, it should be noted that, do not have restrictive requirement to positions such as the starting point of one dimension character gesture and terminals, the characteristic as " hand-written " determines, the result of text identification is only relevant with the pattern of gesture with the process of gesture.
In one example, the one dimension character gesture that character is corresponding corresponds to single stroke, and single stroke is single continuous print stroke, makes in process in one dimension character gesture, and user's finger or instrument of drawing do not leave user's inputting interface.
Detecting unit 120 is configured to the one dimension character gesture that detection user makes on user's inputting interface, and transfer this one dimension character gesture to one-dimensional signal, described one-dimensional signal is the signal only having value in a coordinate system in a dimension.Detecting unit 120 such as comprises the device of pressure transducer and analog to digital converter, and the one dimension character gesture for user being inputted is converted to the one-dimensional signal that computing equipment can process,
Character mother plate database 130 is configured to the template storing each character, and the template of each character corresponds to one dimension stroke, is the set of the one or more sub-stroke according to predefined procedure, and each sub-stroke is line segment substantially point-blank.Detailed introduction is given below by the example of Character mother plate.
Recognition unit 140 is configured to receive this one-dimensional signal from detecting unit, this one-dimensional signal is converted to pending one dimension stroke, this one dimension stroke recognition is respective symbols by the template of each character stored in the one dimension stroke pending based on this and Character mother plate database.
Preferably, handwriting recongnition equipment also comprises display unit 150, the order that may be used for the time sequencing of the sub-stroke that gesture hand-written with one dimension comprises corresponding shows each sub-stroke with spatial order, and/or the text that Identification display goes out (character and/or word), and/or when needs user selects in candidate show candidate.
Two, Character mother plate---one dimension character gesture
Inventor carries out Character mother plate write experiment through choosing a large amount of participant, from being easy to memory, be easy to study, correct rate for input is high, input efficiency much higher aspect is considered, devises various characters template.
Fig. 2 shows the sub-stroke pie graph of each character of Character mother plate according to an embodiment of the invention.
In Fig. 2, for each character, the sub-stroke forming it is pressed sequential write and launches, the set of this little stroke constitutes the one dimension character gesture of corresponding character.
Visible as shown in Figure 2, sub-stroke has direction and length, has three kinds of length, short, medium, length three kinds.Arrow in figure represents the direction of every sub-stroke.The one dimension character gesture of a character be the hand-written gesture design of two-dimentional character in simulation actual characters writing process out, compared to the design based on coding, there is based on hand-written one dimension character gesture design the advantage being easy to identify and be easy to remember.Such as, character a is first descending up descending more again when writing, and therefore corresponding sub-stroke is three descending, up, up sub-strokes, and three gesture strokes when considering actual writing are close, therefore design these three sub-stroke length equal.In addition, for character like the projection-type that some two dimension is write on one dimension, in order to distinguish, carry out little adjustment, such as at gesture ending affix short stroke or " point ", such as, for character " q " and " y ", the projection-type on its one dimension seemingly, descending in being all, in up, long descending, added after the gesture of character q for this reason one all up; Such as character " o " and " v " again, has added one " point " and has distinguished after the gesture of character v.
Finding in the text event detection of the one dimension character gesture shown in Fig. 2 on intelligent glasses being put into practice by many people, being difficult to remember for carrying out distinguishing little adjustment (the short stroke of gesture ending place or the point) user carried out for above-mentioned.In addition, from input accuracy angle, user is difficult to differentiation three seed stroke length.Fig. 3 shows the distribution of the sub-stroke of short, medium, and long in the actual gesture made of user after normalization.Visible, there is more overlap between three seed stroke length.In addition, user makes little delineating (hook) the beginning and end of stroke is unconscious once in a while, and the recognition accuracy of one dimension character gesture is reduced.
For the problems referred to above, the design of inventor to one dimension character gesture is optimized, comprise following in one or more: no longer carry out above-mentioned little adjustment at the end of gesture and (namely no longer add short stroke or " point " at gesture end,), same one dimension character gesture is namely allowed to correspond to a more than character, or on the contrary, multiple character corresponds to same one dimension character gesture, and (character q and y in the Fig. 4 such as described below corresponds to same gesture, e and z, c and i etc.); Only design two kinds of stroke length; Remove the sub-stroke that some are tiny, such as traditional n, m and p (lowercase versions) handwritten form starts with a backward little stroke (the little stroke towards character main body), the handwritten form of traditional d terminates with a little stroke (away from character main body) backward, in order to simplify and strengthen availability, from the one dimension character gesture of these characters, eliminate these little strokes.
The sub-stroke that Fig. 4 shows each character of Character mother plate according to another embodiment of the present invention forms schematic diagram.Only indicate the conventional two-dimensional handwritten stroke of this letter with the straight-line segment of the band arrow of each letter or segment of curve in Fig. 4, circle wherein represents hand-written starting point, the right side of each letter shows corresponding one dimension character gesture, it should be the reciprocal paddling on a linear axis, but be easy to understand and visual, the direction vertical with linear axis is launched.
The design of the Character mother plate set of Fig. 4 is mainly considered to instruct rule as follows:
(1) the hand-written gesture of conventional two-dimensional is simulated, so that easy learning and memory;
(2) minimize the length of the sub-stroke wherein contained, make user can make character gesture accurately and efficiently;
(3) number of character neutron stroke is minimized, so that execution character gesture efficiently;
(4) one stroke input, single stroke refers to single continuous print stroke, namely makes in process in the one dimension character gesture of a character, and user's finger or instrument of drawing do not leave user's inputting interface, so that user's execution character gesture efficiently.
As shown in Figure 4, two seed stroke length are had at most in each one dimension character gesture.Devise 13 kinds of character gestures, be mapped to 26 letters, for keeping succinct efficient, each one dimension character gesture can not be associated with more than four characters.For a character, if map directly to the one-dimensional space there is not preferred mapping, different process has been carried out according to the character of character, such as, for letter " e ", " z " and " s ", different from other letters, not directly carry out two-dimentional hand-written gesture to the mapping in the one-dimensional space, but 90-degree rotation first left, and then by hand-written for two dimension hand modelling to the one-dimensional space; For alphabetical x, there are two sub-strokes, the gesture of alphabetical n is distributed to it; In addition as previously mentioned, alphabetical n, m and p (lowercase versions) handwritten form is started with a backward little stroke (the little stroke towards character main body), the handwritten form of traditional d terminates with a little stroke (away from character main body) backward, in order to simplify and strengthen availability, from the one dimension character gesture of these characters, eliminate these little strokes.
Character mother plate database is above the preferred embodiment of the present invention, but not as display, can change as required, replaces, increases or delete Character mother plate.
Three, one dimension hand-written character input method
Writing words input can comprise character input and words enter, and corresponding Text region also comprises character recognition and vocabulary identification, and wherein character input/identification is the basis of words enter/identification.Description below first describes character recognition, follow-uply describes vocabulary identification again.
Below with reference to the accompanying drawings 5 describe the one dimension hand-written character input method 200 performed by one dimension handwriting input equipment according to the embodiment of the present invention, one dimension handwriting input equipment comprises user's inputting interface and Character mother plate database.
In step S210, manually one dimension character gesture is made in chronological order in the mode contacting user's inputting interface with finger or instrument of drawing in response to user, detect the one dimension character gesture that user makes on user's inputting interface, transfer this one dimension character gesture to one-dimensional signal, described one-dimensional signal is the signal only having value in a coordinate system in a dimension, and so-called one dimension character gesture is reciprocal paddling point-blank substantially.
In step S220, receive this one-dimensional signal from detecting unit, this one-dimensional signal is converted to pending one dimension stroke, the template of each character stored in the one dimension stroke pending based on this and Character mother plate database, be respective symbols by this one dimension stroke recognition, wherein the template of each character corresponds to one dimension stroke, is the set of the one or more sub-stroke according to predefined procedure, and each sub-stroke is line segment substantially point-blank.
Here, Character mother plate database can adopt the Character mother plate database such as described with reference to accompanying drawing 4.
In one example, one dimension handwriting input equipment also comprises display unit.Described display unit can be such as LCD screen, also can be wear-type miniature display screen (the wear-type miniature display screen of such as Google, can by projection formed virtual screen) etc.
Handwritten text input method also comprises: by means of display unit, the order corresponding with the time sequencing of the sub-stroke that the hand-written gesture of one dimension comprises shows each sub-stroke with spatial order, and the text that Identification display goes out (character and/or word), and when needs user selects in candidate show candidate.
In one example, described one dimension handwriting input equipment can work in character input mode or words enter pattern.Such as Text region equipment can be given tacit consent to and works in words enter pattern, and user can be switched to character input mode by performing input pattern blocked operation.Exemplarily, input pattern blocked operation can be finger residence time on user's inputting interface exceed predetermined threshold, and such as, more than 300 milliseconds, namely such user can switch without the need to lifting finger between character input mode and words enter pattern.As required, other input pattern blocked operation can be designed.
Fig. 6 shows the example of the graphic user interface according to the embodiment of the present invention, and when one dimension handwriting input equipment is Google glasses, this graphic user interface may be displayed on the virtual screen of Google glasses.
Graphic user interface 300 is divided into three regions: text filed 310, character zone 320 and vocabulary region 330.Text filed 310 display input texts, the character candidates that character zone 320 Identification display goes out, the word candidate of vocabulary region 330 Identification display.
In addition, in one example, also feed back at the two dimensional visible of described text filed display one dimension character gesture, wherein said text filed height mapping to user's inputting interface one dimension character gesture one-dimensional square whole length upwards (such as, text filed height corresponds to the tangible length range of intelligent glasses temple), described text filed horizontal direction represents time shaft, described hand-written recognition method also comprises: in described text filed rise time axle display one dimension character gesture, thus provide the two dimensional visible of one dimension character gesture to feed back.Gesture (namely lower stroke, upper stroke, lower stroke) such as shown in Fig. 6 is corresponding to character u, s and a, therefore text filed in the middle of 310 by lower stroke in this one-dimensional space, on draw and lower stroke show (shape is similar to capitalization " N ") according to time sequencing expansion, this expansion display can be superimposed upon text filed for translucent form, the two dimensional visible feedback guidance of input stroke is provided thus, simultaneously at character zone 320 display candidate characters " u ", " s " and " a " during user's input; And in vocabulary region 330 show candidate vocabulary " is ", " la ", " can ", " last " and " late ", and highlighted display first candidate " is ", for the display brightness of the character inputted in candidate's vocabulary higher than the display brightness (such as the display brightness of character " c " and " a " is greater than the display brightness of character " n " in candidate's vocabulary " can ") of other characters of prediction.Now, detect in response to detecting unit the selection gesture that user makes, such as, refer to pat user's inputting interface with two, and select current highlighted vocabulary (" is ").Alternatively, detect in response to detecting unit the mobile gesture that user makes, such as, two refer to move on user's inputting interface, and processor switches the character candidates (such as moving to next candidate " la ") be highlighted.
In one example, after have selected a vocabulary, one dimension handwriting input equipment adds a space automatically after vocabulary.
In one example, described one dimension hand-written character input method also comprises: detect that gesture separated in the character that user makes in response to detecting unit, continuous print one dimension character gesture is distinguished, to terminate the input of last character, and prepares the input receiving next character.Character separation gesture can exceed predetermined threshold for the finger residence time, clicks, press down etc. in one example.
In one example, one dimension character gesture corresponds to single stroke, and single stroke is single continuous print stroke, makes in process in one dimension character gesture, and user's finger or instrument of drawing do not leave user's inputting interface, improve the efficiency of character input thus.
In one example, the sub-stroke of the stroke of the template of each character has length and direction, and wherein length is one of two kinds of predetermined lengths.As previously mentioned, the accuracy and efficiency of input can be improved like this.
In one example, described one dimension character gesture and the hand-written tendency similar process of reality two dimension character are to the projection of the hand-written gesture of two-dimentional character in the one-dimensional space.
Below with reference to Fig. 7, the method 400 one-dimensional signal corresponding for one dimension character gesture being converted to pending one dimension stroke according to the embodiment of the present invention is described.
In step 410, identify flex point, flex point refers to carry out in time in this one dimension stroke forming process the point that stroke direct of travel changes, and flex point comprises starting point and the end point of stroke.
At step 420 which, the flex point being less than intended pixel length the flex point middle distance time in front flex point is removed.
This step is remove the flex point as noise.According to the design of Character mother plate, the distance between two flex points should be not too short, if therefore the spacing of two adjacent comers is less than predetermined threshold, such as, is less than 20 pixels, then can be noise depending on flex point below, be removed.
In step 430, line segment is divided based on remaining flex point.
That is, on the time, namely two adjacent flex points form line segment, are also the sub-stroke of candidate.
In step 440, get rid of length in head and the tail line segment and be less than the line segment of predetermined threshold, residue line segment is normalized, obtains the sequence of set as the sub-stroke according to time sequencing of the line segment after normalization.
As previously mentioned, the length in head and the tail line segment of getting rid of in this step is less than the line segment of predetermined threshold, is intended to remove user and makes little delineating (hook) the beginning and end of stroke is unconscious once in a while.
Here being normalized line segment can by being undertaken the length of the length of each line segment divided by the longest sub-stroke.Here the longest sub-stroke can be line segment the longest in such as each line segment above-mentioned.
Through above-mentioned process, obtain the set of line segment, namely according to the sequence of sub-stroke of time sequencing arrangement, the datumization of the one dimension character gesture (being also referred to as one dimension stroke herein) made as user represents.
After obtaining one dimension stroke, as shown in step S220 in Fig. 2, this one dimension stroke recognition is respective symbols by the template of each character stored in the one dimension stroke pending based on this and Character mother plate database
Various machine learning method can be utilized to carry out this identification, such as template matching method, as by directly in feature space, calculate the distance between one dimension stroke characteristic of correspondence vector and the template characteristic of correspondence vector of each character, select one as recognition result (or selecting multiple as recognition result candidate).
According to one embodiment of the invention, can utilize bayes method to try to achieve user do one dimension character gesture and correspond to the probability of certain character.
Fig. 8 shows the template based on each character stored in pending one dimension stroke and Character mother plate database, is the process flow diagram of the method 500 of respective symbols by this one dimension stroke recognition.
In the method shown in Fig. 8, suppose that the sub-number of strokes comprised in one dimension stroke to be identified is n, the template only equaling the character of n equally for comprised sub-number of strokes carrys out calculating probability, and sub-number of strokes is not equal to the template of the character of n, think that one dimension stroke to be identified be the probability of this character is zero.
As shown in Figure 8, in step S510, determine the number of the sub-stroke that this one dimension stroke comprises.Then, step S520 is proceeded to.
In step S520, searching character template database, determines that template neutron number of strokes equals the template of the number of the sub-stroke that described this one dimension stroke determined comprises.Then, step S530 is proceeded to.
In step S530, each in determined template, according to formation order, determines that every sub-stroke in this template is rendered as the probability of the sub-stroke of correspondence in described one dimension stroke, calculates the probability that described one dimension stroke corresponds to this template.
If pending one dimension stroke is S, the sub-stroke forming it is s 1, s 2s nif 26 Character mother plates are T [1], T [2] ... T [j] ... T [26], the sub-stroke forming Character mother plate T [j] is T [j] 1, T [j] 2t [j] n.Then need to calculate T [j].According to Bayesian formula, below formula (1) set up,
P(T[j]/S)＝P(T[j],S)/P(S)
＝(P(T[j])×P(S/T[j]))/P(S)(1)
Probability P (S) is total for the calculating of the posterior probability P (T [j]/S) of each template, therefore on impact be identical, without the need to considering this.The prior probability that certain character of probability P (T) i.e. occurs can, by adding up acquisition in advance, namely be known.In one example, for simplifying, can think that the probability of occurrence of each character is equal.We think that between each sub-stroke be independently, then formula (2) is set up
P ( S / T [ j ] ) = Π i = 1 n P ( S i / T [ j ] i ) - - - ( 2 )
Wherein probability P (S i/ T [j] i), i.e. i-th sub-stroke T [j] of template T [j] ibe rendered as the sub-stroke S in user's one dimension character gesture S iprobability, by adding up acquisition in advance, can calculate the probability P (T [j]/S) that this one dimension character gesture when user makes one dimension character gesture S represents character T [j] thus.
Resulting in the probability that one dimension stroke corresponds to each template.
After step S530 completes, proceed to step S540.
In step S540, corresponding to the probability of each template based on described one dimension stroke, is corresponding character by described one dimension stroke recognition.
Such as, determining the template that maximum probability is corresponding, is this template by described one dimension stroke recognition, in other words, identifies the character that one dimension character gesture that user does is corresponding.
In one example, also comprise display unit, the display screen of this display unit be (necessary being, or virtual display screen) such as configure as shown in Figure 6, comprising character zone 320 and text filed 310, text filed 310 display input texts, for when there is multiple character candidates in character zone 320, a display predetermined number character candidates, such as, shows 3 character candidates in Fig. 6 in character zone 320.
In one example, character identifying method also comprises: in response to detecting that user makes one dimension character gesture and terminates to determine gesture, show the determined character candidates corresponding with this one dimension character gesture, described one dimension character gesture terminates to determine that gesture instruction one dimension character gesture terminates.In one example, one dimension character gesture terminates to determine that gesture can be pause gesture, such as, for intelligent glasses, as on the temple side frame of user's inputting interface, keeps that finger is motionless reaches the schedule time.
Next, in response to detecting that user makes character and selects gesture, select the current candidate characters be highlighted, and be presented at the viewing area of the text of indicative input, or in response to detecting that user makes candidate characters and switches gesture, switching the current candidate characters that is highlighted and being switched to highlighted candidate characters with highlighted fashion display.In one example, in default situations, the candidate characters acquiescence that such as probable value of aforementioned calculating is the highest is as highlighted candidate characters, highlighted.Candidate characters switches gesture for such as to point mobile gesture, often can move preset distance, be then switched to highlighted candidate characters along the candidate characters that moving direction is adjacent with current highlighted candidate characters.In one example, character selects gesture can lift for finger the action leaving user's inputting interface.
In brief, under intelligent glasses environment, the operation scenario of character input is, finger paddling on leg of spectacles, at the end of the one dimension finger that character is corresponding, user points and keeps motionless, user can observe each candidate characters shown on the display screen simultaneously, then highlight characters is switched by moveable finger, finally by lifting out finger to choose a character, thus in this Charactes Display within display screen text filed.
In some cases, at the end of one dimension character gesture, finger is in the end of user's inputting interface, such as still for intelligent glasses, one dimension character gesture finger is the paddling along leg of spectacles longitudinal direction, now probably a character end of input time finger rest on temple two ends (for describe direction, be referred to as hold A and end B) in end A, at this moment again along with stop before moving direction (namely to hold A or from the direction of hold A movement) move and have difficulties, and to holding the finger paddling in B direction to be easy.For this reason, the DISPLAY ORDER of character candidates can be made for ease of along from end A to the order of carrying out character switching during the direction paddling finger of end B, namely first highlighted display is that probability is the highest, each candidate characters presses probability order display from high to low on the display screen, when to end B moveable finger, the character of secondary high probability is highlighted.
In one example, can delete a character by character deletion gesture, character deletion gesture such as upward sliding or can press button of taking pictures.Still for intelligent glasses, test discovery through multiple user, user more preference presses this gesture of button of taking pictures, and this may be because the upward sliding gesture in temple is not be easy to carry out.
Described above is and how to utilize bayes method identification character.Under being described in words enter pattern below with reference to Fig. 9, utilize the illustrative methods 600 of the vocabulary of bayes method and language model prediction input.
In one example, described one dimension handwriting input equipment also comprises language model data storehouse, stores the probability distribution of instruction by character, assigns probability to the information of the word be made up of character.It should be noted that, comprising here, both comprised the situation that language model data stock is stored in one dimension handwriting input equipment this locality, also contain language model data storehouse and be distributed in situation on remote equipment.
As shown in Figure 9, in step S610, under words enter pattern, recognition unit reference language model database calculates the probability inputting specific vocabulary under given one dimension stroke sequence condition in real time.
In this example, split between each one dimension stroke, such as, split by character separator.
If the list entries of k one dimension stroke (corresponding to a character) is I, be k one dimension stroke S [1], S [2] ... S [i] ... the sequence of S [k], each one dimension stroke corresponds to a character.If a vocabulary W has k letter, i.e. W=L 1l il k, as the list entries I of given one dimension stroke, input be that the probability of vocabulary W can represent with formula (3)
P(W/I)＝P(W,I)/P(I)(3)
Here, P (I) is total item when calculating all vocabulary W posterior probability, can not consider.Joint probability P (W, I) can represent with formula (4)
P(W,I)＝P(W)×P(I/W)(4)
Then we suppose that the input of each character is independent of one another, then P (W, I) can be converted to formula (5)
P ( W , I ) = P ( W ) × Π i = 1 k P ( S [ i ] / L i ) - - - ( 5 )
With T (L i) represent alphabetical L itemplate can obtain formula (6)
P ( W , I ) = P ( W ) × Π i = 1 k P ( S [ i ] / T ( L i ) ) - - - ( 6 )
And P (S [i]/T (L i) formula above (2) can be utilized to obtain.
The posterior probability of the vocabulary identical with the character number of input character sequence can be calculated thus, if the vocabulary number deficiency that the vocabulary not finding number of letters to mate or the number of letters found are mated, the posterior probability of the vocabulary with more golygram can be calculated, this generally needs reference language model, the article being entitled as " Languagemodelingforsoftkeyboard " that the document that pertinent reference language model calculates vocabulary can be delivered on IUI ' 02 194-195 page with reference to Goodman etc., it will not go into details here.
In step S620, after calculating the probability of one dimension stroke sequence corresponding to each vocabulary, based on the probability of each vocabulary of input, candidate's vocabulary is sorted, and display on the display unit, select for user, such as, vocabulary region 330 shown in Fig. 6.
One dimension hand-written character input method according to the embodiment of the present invention and equipment have been described with reference to the drawings above, provide a kind of be particularly suitable for carrying out inputting in the relatively long and narrow one dimension input space character input method and equipment.
According to the preferred embodiment of the present invention, provide the means that single character represents with single stroke, thus provide efficient input mode.
According to the preferred embodiment of the present invention, by the simulation of two-dimentional character writing gesture and the determination of antithetical phrase stroke length kind, devise and be convenient to study, efficiently one dimension hand writing gesture, provide one efficiently, character input method and input equipment accurately.
Aforementioned description is only illustrative, is intended to pass on invention theory of the present invention to those skilled in the art in complete and understandable mode.Aforementioned description should as restriction of the present invention, and those skilled in the art can select some details, means, device etc. on invention ideal basis of the present invention, change or supplement.
In description above, one dimension handwriting input equipment and the method that example describes the embodiment of the present invention is input as with English alphabet and word, but one dimension handwriting input equipment of the present invention and method are also suitable for the input of other spoken and written languages, as long as can by hand-written for two dimension hand modelling to the hand-written gesture of one dimension.
In addition, in example above, it is all the application illustrating the embodiment of the present invention in the input space of approximate one dimension, the embodiment of the present invention is also applicable to the input space of two dimension certainly in fact, such as on the touch-screen etc. of current handset touch panel, notebook computer, also can carry out one dimension character gesture (now subsequent detection unit also only extracts one-dimensional signal corresponding to one dimension character gesture), the use of such as screen display can be left in more space by this.
In addition, here for finger or draw instrument and user and take in interracial contact the form that user inputs is described, in fact, the present invention is not limited thereto, also user one dimension character gesture can be carried out in a non contact fashion, the one dimension character gesture of reciprocal paddling made by the laser pen of the hand-held Emission Lasers of such as user, as long as detecting unit can detect such laser trace, also just can obtain the input of user.For another example, user makes the one dimension character gesture of reciprocal paddling aloft with finger, and detecting unit such as identifies such gesture by image procossing, then also can the acquisition of completing user handwriting input.
The main example application using intelligent glasses as input method of the present invention above, but this is only example, input method of the present invention can be applied to the equipment such as intelligent watch, Intelligent bracelet, smart mobile phone.
In aforementioned description, the one dimension character gesture of user is made with finger, but this is only example, can select other body parts of user as required, such as toe, wrist etc.In addition, also can be unfavorable for human body, and utilize external tool to input, such as writing pencil etc.
According to another embodiment of the present invention, provide a kind of one dimension handwriting input equipment, comprising: user's inputting interface, manually make one dimension character gesture in chronological order in the mode contacting this user's inputting interface for user with body part or instrument; Detecting unit, be configured to the one dimension character gesture that detection user makes on user's inputting interface, transfer this one dimension character gesture to one-dimensional signal, described one-dimensional signal is the signal only having value in a coordinate system in a dimension; Character mother plate database, is configured to the template storing each character, and the template of each character is the proper vector of described one-dimensional signal being carried out to statistically feature extraction acquisition; Recognition unit, is configured to receive from this one-dimensional signal of detecting unit, this one-dimensional signal is converted to pending proper vector, and the template of each character stored in the proper vector pending based on this and Character mother plate database, identifies and obtain respective symbols.
According to an embodiment, one dimension character gesture can be one of following items: the reciprocal paddling substantially point-blank carried out on described user's inputting interface; The single-point that described user's inputting interface carries out presses down; And the rotation on the one dimension angle coordinate system carried out on described user's inputting interface.
According to another embodiment of the present invention, provide a kind of one dimension handwriting input equipment, comprise: user's input detection unit, obtain the one dimension character gesture that user makes in chronological order, so-called one dimension character gesture is reciprocal paddling point-blank substantially, and transferring this one dimension character gesture to one-dimensional signal, described one-dimensional signal is the signal only having value in a coordinate system in a dimension; Character mother plate database, is configured to the template storing each character, and the template of each character corresponds to one dimension stroke, is the set of the one or more sub-stroke according to predefined procedure, and each sub-stroke is line segment substantially point-blank; Recognition unit, configuration receives this one-dimensional signal from detecting unit, this one-dimensional signal is converted to pending one dimension stroke, and this one dimension stroke recognition is respective symbols by the template of each character stored in the one dimension stroke pending based on this and Character mother plate database.
Described one dimension character gesture can be that user's body position or instrument of using are made on user's inputting interface with the way of contact, also can make in a non contact fashion, described user's detecting unit can comprise image taking and treating apparatus, pressure-detecting device, infrared detecting device etc.
Be described above various embodiments of the present invention, above-mentioned explanation is exemplary, and non-exclusive, and be also not limited to disclosed each embodiment.When not departing from the scope and spirit of illustrated each embodiment, many modifications and changes are all apparent for those skilled in the art.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claim.

Claims (10)

1. an one dimension handwriting input equipment, comprising:
User's inputting interface, manually makes one dimension character gesture in the mode contacting this user's inputting interface for user in chronological order with body part or instrument of drawing, and so-called one dimension character gesture is reciprocal paddling point-blank substantially;
Detecting unit, be configured to the one dimension character gesture that detection user makes on user's inputting interface, transfer this one dimension character gesture to one-dimensional signal, described one-dimensional signal is the signal only having value in a coordinate system in a dimension;
Character mother plate database, is configured to the template storing each character, and the template of each character corresponds to one dimension stroke, is the set of the one or more sub-stroke according to predefined procedure, and each sub-stroke is line segment substantially point-blank;
Recognition unit, configuration receives this one-dimensional signal from detecting unit, this one-dimensional signal is converted to pending one dimension stroke, and this one dimension stroke recognition is respective symbols by the template of each character stored in the one dimension stroke pending based on this and Character mother plate database.
2. one dimension handwriting input equipment according to claim 1, also comprises:
Display unit;
Wherein, at least one Character mother plate stored in Character mother plate database corresponds to multiple character,
Wherein, when processor identification one dimension stroke corresponds to multiple character as character candidates, processor performs process and makes to show the plurality of character candidates over the display, and wherein one of multiple character candidates is highlighted at any one time; And
Detect in response to detecting unit the selection gesture that user makes, the current character candidates be highlighted of processor selection is as character identification result; And detecting in response to detecting unit the mobile gesture that user makes, processor switches the character candidates be highlighted.
3. according to the one dimension handwriting input equipment of claim 1 or 2, detect that gesture separated in the character that user makes in response to detecting unit, continuous print one dimension character gesture distinguishes by processor, to terminate the input of last character, and prepares the input receiving next character.
4. one dimension handwriting input equipment according to claim 3, gesture separated in described character is pause gesture, and pause gesture corresponds to user's finger or instrument of drawing dead time on user's inputting interface exceedes predetermined threshold.
5. according to the one dimension handwriting input equipment of claim 1 or 2, so-called one dimension character gesture corresponds to single stroke, single stroke is single continuous print stroke, makes in process in one dimension character gesture, and user's finger or instrument of drawing do not leave user's inputting interface.
6. according to the one dimension handwriting input equipment of claim 1 or 2, the sub-stroke of the stroke of the template of each character has length and direction, and wherein length is one of two kinds of predetermined lengths.
7. according to the one dimension handwriting input equipment of claim 1 or 2, described one dimension character gesture and the hand-written tendency similar process of reality two dimension character.
8. an one dimension hand-written character input method for one dimension handwriting input equipment execution, one dimension handwriting input equipment comprises user's inputting interface and Character mother plate database, and described one dimension hand-written character input method comprises:
Manually one dimension character gesture is made in chronological order in the mode contacting user's inputting interface with finger or instrument of drawing in response to user, detect the one dimension character gesture that user makes on user's inputting interface, transfer this one dimension character gesture to one-dimensional signal, described one-dimensional signal is the signal only having value in a coordinate system in a dimension, and so-called one dimension character gesture is reciprocal paddling point-blank substantially;
Receive this one-dimensional signal from detecting unit, this one-dimensional signal is converted to pending one dimension stroke, the template of each character stored in the one dimension stroke pending based on this and Character mother plate database, be respective symbols by this one dimension stroke recognition, wherein the template of each character corresponds to one dimension stroke, for the set of the one or more sub-stroke according to predefined procedure, each sub-stroke is line segment substantially point-blank.
9. an one dimension handwriting input equipment, comprising:
User's inputting interface, manually makes one dimension character gesture in the mode contacting this user's inputting interface for user in chronological order with body part or instrument;
Detecting unit, be configured to the one dimension character gesture that detection user makes on user's inputting interface, transfer this one dimension character gesture to one-dimensional signal, described one-dimensional signal is the signal only having value in a coordinate system in a dimension;
Character mother plate database, is configured to the template storing each character, and the template of each character is the proper vector of described one-dimensional signal being carried out to statistically feature extraction acquisition,
Processor, configuration receives from this one-dimensional signal of detecting unit, and this one-dimensional signal is converted to pending proper vector, and the template of each character stored in the proper vector pending based on this and Character mother plate database, identifies and obtain respective symbols.
10. one dimension handwriting input equipment according to claim 9, described one dimension character gesture is one of following items:
The reciprocal paddling substantially point-blank that described user's inputting interface carries out;
The single-point that described user's inputting interface carries out presses down; And
Rotation on the one dimension angle coordinate system that described user's inputting interface carries out.
CN201511009156.7A 2015-12-29 2015-12-29 One-dimensional handwriting input equipment and one-dimensional hand-written character input method Active CN105549890B (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201511009156.7A CN105549890B (en) 2015-12-29 2015-12-29 One-dimensional handwriting input equipment and one-dimensional hand-written character input method
PCT/CN2016/105694 WO2017114002A1 (en) 2015-12-29 2016-11-14 Device and method for inputting one-dimensional handwritten text

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201511009156.7A CN105549890B (en) 2015-12-29 2015-12-29 One-dimensional handwriting input equipment and one-dimensional hand-written character input method

Publications (2)

Publication Number Publication Date
CN105549890A true CN105549890A (en) 2016-05-04
CN105549890B CN105549890B (en) 2019-03-05

Family

ID=55829095

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201511009156.7A Active CN105549890B (en) 2015-12-29 2015-12-29 One-dimensional handwriting input equipment and one-dimensional hand-written character input method

Country Status (2)

Country Link
CN (1) CN105549890B (en)
WO (1) WO2017114002A1 (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106598268A (en) * 2016-11-10 2017-04-26 清华大学 Text input method and electronic device
WO2017114002A1 (en) * 2015-12-29 2017-07-06 清华大学 Device and method for inputting one-dimensional handwritten text
CN109920309A (en) * 2019-01-16 2019-06-21 深圳壹账通智能科技有限公司 Sign language conversion method, device, storage medium and terminal
CN110377914A (en) * 2019-07-25 2019-10-25 腾讯科技（深圳）有限公司 Character identifying method, device and storage medium
CN112861709A (en) * 2021-02-05 2021-05-28 金陵科技学院 Hand-drawn sketch recognition method based on simple strokes
CN113711232A (en) * 2019-03-20 2021-11-26 微软技术许可有限责任公司 Object detection and segmentation for inking applications

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111506185B (en) * 2019-01-31 2023-09-29 珠海金山办公软件有限公司 Method, device, electronic equipment and storage medium for operating document
CN111046751B (en) * 2019-11-22 2024-02-13 华中师范大学 Formula identification method and device

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101739118A (en) * 2008-11-06 2010-06-16 大同大学 Video handwriting character inputting device and method thereof
US8199126B1 (en) * 2011-07-18 2012-06-12 Google Inc. Use of potential-touch detection to improve responsiveness of devices
US8319746B1 (en) * 2011-07-22 2012-11-27 Google Inc. Systems and methods for removing electrical noise from a touchpad signal
CN104063069A (en) * 2014-07-03 2014-09-24 南京吉隆光纤通信股份有限公司 Character input device convenient to use
CN104133559A (en) * 2014-07-04 2014-11-05 浙江大学 Display method for candidate words input from touch screen

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1845040A (en) * 2006-03-21 2006-10-11 北京三桥科技有限公司 Single-key input terminal and use method thereof
CN101706689B (en) * 2009-11-25 2013-03-13 福州福昕软件开发有限公司 Method and device for inputting characters via direction key
CN102314252B (en) * 2010-06-30 2013-09-11 汉王科技股份有限公司 Character segmentation method and device for handwritten character string
US9563355B2 (en) * 2013-04-22 2017-02-07 Rajeev Jain Method and system of data entry on a virtual interface
JP2015148946A (en) * 2014-02-06 2015-08-20 ソニー株式会社 Information processing device, information processing method, and program
CN105549890B (en) * 2015-12-29 2019-03-05 清华大学 One-dimensional handwriting input equipment and one-dimensional hand-written character input method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101739118A (en) * 2008-11-06 2010-06-16 大同大学 Video handwriting character inputting device and method thereof
US8199126B1 (en) * 2011-07-18 2012-06-12 Google Inc. Use of potential-touch detection to improve responsiveness of devices
US8319746B1 (en) * 2011-07-22 2012-11-27 Google Inc. Systems and methods for removing electrical noise from a touchpad signal
CN104063069A (en) * 2014-07-03 2014-09-24 南京吉隆光纤通信股份有限公司 Character input device convenient to use
CN104133559A (en) * 2014-07-04 2014-11-05 浙江大学 Display method for candidate words input from touch screen

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017114002A1 (en) * 2015-12-29 2017-07-06 清华大学 Device and method for inputting one-dimensional handwritten text
CN106598268A (en) * 2016-11-10 2017-04-26 清华大学 Text input method and electronic device
CN106598268B (en) * 2016-11-10 2019-01-11 清华大学 Text entry method and electronic equipment
CN109920309A (en) * 2019-01-16 2019-06-21 深圳壹账通智能科技有限公司 Sign language conversion method, device, storage medium and terminal
CN109920309B (en) * 2019-01-16 2023-02-03 深圳壹账通智能科技有限公司 Sign language conversion method, device, storage medium and terminal
CN113711232A (en) * 2019-03-20 2021-11-26 微软技术许可有限责任公司 Object detection and segmentation for inking applications
CN110377914A (en) * 2019-07-25 2019-10-25 腾讯科技（深圳）有限公司 Character identifying method, device and storage medium
CN112861709A (en) * 2021-02-05 2021-05-28 金陵科技学院 Hand-drawn sketch recognition method based on simple strokes

Also Published As

Publication number Publication date
WO2017114002A1 (en) 2017-07-06
CN105549890B (en) 2019-03-05

Similar Documents

Publication Publication Date Title
CN105549890A (en) One-dimensional handwritten character input equipment and one-dimensional handwritten character input equipment
US6795579B2 (en) Method and apparatus for recognizing handwritten chinese characters
US9025879B2 (en) Electronic apparatus and handwritten document processing method
US7487461B2 (en) System and method for issuing commands based on pen motions on a graphical keyboard
CN108700996B (en) System and method for multiple input management
CN102449640B (en) Recognizing handwritten words
US20060119588A1 (en) Apparatus and method of processing information input using a touchpad
US9606981B2 (en) Electronic apparatus and method
US8341556B2 (en) Method and system for attention-free user input on a computing device
US20150169948A1 (en) Electronic apparatus and method
CN109074224A (en) For the method for insertion character and corresponding digital device in character string
EP2626813B1 (en) Apparatus and method for guiding handwriting input for handwriting recognition
CN104903836A (en) Method and device for typing on mobile computing devices
JP2019220155A (en) Handwriting input display device, handwriting input display method and program
KR102347554B1 (en) Systems and methods for beautifying digital ink
WO2013171919A1 (en) Display control device, control program and display device control method
US9940536B2 (en) Electronic apparatus and method
KR20210061523A (en) Electronic device and operating method for converting from handwriting input to text
US11514696B2 (en) Display device, display method, and computer-readable recording medium
JP5897726B2 (en) User interface device, user interface method, program, and computer-readable information storage medium
US10127478B2 (en) Electronic apparatus and method
US11087156B2 (en) Method and device for displaying handwriting-based entry
US11244138B2 (en) Hologram-based character recognition method and apparatus
WO2016031016A1 (en) Electronic device, method, and program
Tonouchi et al. Text input system using online overlapped handwriting recognition for mobile devices

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant