CN103080878B - One or more groups of method and apparatus are divided into for hand-written stroke will be overlapped - Google PatentsOne or more groups of method and apparatus are divided into for hand-written stroke will be overlapped Download PDF
- Publication number
- CN103080878B CN103080878B CN201080068735.8A CN201080068735A CN103080878B CN 103080878 B CN103080878 B CN 103080878B CN 201080068735 A CN201080068735 A CN 201080068735A CN 103080878 B CN103080878 B CN 103080878B
- Prior art keywords
- Prior art date
- 230000011218 segmentation Effects 0.000 claims abstract description 44
- 238000004590 computer program Methods 0.000 claims abstract description 24
- 238000000034 methods Methods 0.000 claims description 11
- 238000010606 normalization Methods 0.000 claims description 11
- 230000000875 corresponding Effects 0.000 claims description 7
- 280000706878 Group It companies 0.000 claims description 4
- 238000005516 engineering processes Methods 0.000 description 12
- 238000004891 communication Methods 0.000 description 8
- 230000015654 memory Effects 0.000 description 5
- 238000004458 analytical methods Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000010586 diagrams Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 241000272201 Columbiformes Species 0.000 description 2
- 230000001413 cellular Effects 0.000 description 2
- 239000003086 colorants Substances 0.000 description 2
- 229920001690 polydopamine Polymers 0.000 description 2
- 239000007787 solids Substances 0.000 description 2
- 241000256844 Apis mellifera Species 0.000 description 1
- 210000004247 Hand Anatomy 0.000 description 1
- 101710079960 TPPP Proteins 0.000 description 1
- 280000874760 Third Generation companies 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000005314 correlation function Methods 0.000 description 1
- 230000001809 detectable Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 101710043581 endX Proteins 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006011 modification reactions Methods 0.000 description 1
- 230000000051 modifying Effects 0.000 description 1
- 230000001537 neural Effects 0.000 description 1
- 230000003287 optical Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000002269 spontaneous Effects 0.000 description 1
- 230000001360 synchronised Effects 0.000 description 1
- 230000001052 transient Effects 0.000 description 1
- G06—COMPUTING; CALCULATING; COUNTING
- G06K—RECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
- G06K9/00—Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
- G06K9/00852—Recognising whole cursive words
- G06K9/00865—Recognising whole cursive words using stroke segmentation
- G06—COMPUTING; CALCULATING; COUNTING
- G06K—RECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
- G06K9/00—Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
- G06K9/00402—Recognising digital ink, i.e. recognising temporal sequences of handwritten position coordinates
- G06K9/00409—Preprocessing; Feature extraction
- G06K9/00416—Sampling; contour coding; stroke extraction
The illustrative embodiments of the present invention are related to overlap hand-written process, and more particularly, to overlap will be included A series of hand-written strokes are divided into one or more groups of strokes.
Various equipment pass through touch screen or other input equipments are captured or receive handwriting input.For example, many computers, flat Plate computer, mobile phone, personal digital assistant（PDA）Include touch screen with other kinds of electronic equipment, which allows user such as with handss The form write provides sense of touch.Handwriting input can be captured by electronic equipment, and it is processed with attempt to recognize hand-written character with Just the future-action of electronic equipment is informed, including the storage or transmission of the expression of such as handwriting input.
In order that character recognition is easier, handwriting input is typically processed so that handwriting input is divided into kinds of characters, Such as letter, numeral or other characters.There are several different technologies for splitting handwriting input.A kind of technology utilizes each word Time-out between symbol input.According to the technology, such as user interface, touch screen, it is only necessary to write region including one.User can be This is write.Then suspend, next word is thought to receive the character and to remove this and write region The reception of symbol is ready to this and writes region.Therefore, the time-out between each character input is for hand-written segmentation.
Another kind of technology writes region using with two or more, such as the user interface of two or more touch screens.With A character can be write in these write one in region in family, can then switch to other and write region to write a word Symbol.When user just writes a middle written character in region at these, other are written on before and write the character in region Can be received and be processed, and other are write region and then can be cleared in order to other books be got out for the reception of next character Write region.Therefore, the switching that difference is write between region is used for hand-written segmentation.
Yet another technique single writes region using sizable.Writing in region, user can using be comparable to by Character writing mode on a sheet of paper several characters of continuous writing one by one.The technology is using positioned at institute's written character Between space interval and bottom intelligence so as to appropriate Ground Split handwriting input.
In another technology, user interface offer is single to write region, writes in region at this, and user can adopt word The several characters of mode continuous writing overlapped each other without space interval between symbol.By using bottom intelligence, the character of overlap Can be segmented to separate a character and other characters.However, the program is than using the identifying processing for generally adopting above Technology it is more complicated.Under maximal possibility estimation, a series of strokes with errors can be divided into character by the identifying processing.Should Identifying processing has very big calculating intensity, and for real-time operation is particularly mini-plant, very challenging property.
Using the continuous hand-written multiple technologies of overlap, it is allowed to which the electronic equipment with relatively small user interface is received and used The continuous handwriting input of the finger of user or other writing pencils, so that handwriting input can be compared to the input of one at a time character Relatively fast mode is provided.However, with regard to segmentation and user mutual, may challenge to overlapping continuous hand-written dependence. This respect, overlap is continuously hand-written to increase hand-written so that each character and other characters are separated related to effective Ground Split overlap Difficulty, it is particularly continuous hand-written with space interval compared between more common institute's write characters.In addition, multiple characters Overlap and difficulty can be caused to user, because user can not because of the noisy and complicated background that other characters for overlapping are caused Clearly check user is currently writing anything.
The content of the invention
According to illustrative embodiments, there is provided a kind of method, device and computer program, so as to will including overlap A series of hand-written strokes are divided into one or more groups of strokes.The segmentation can be divided into any segmentation of one or more characters Perform before operation, and in fact every group of stroke can be a part for character or character, but there is no one group anyway Stroke includes the stroke from more than one character.By a series of strokes are divided into one or more groups of strokes, subsequently should Serial stroke is divided into one or more characters to become more effective.In addition, one or more groups of strokes before can be relatively failing to understand Show different modes to show, while it is hand-written to allow user clearly to check newest stroke continuously to receive overlap.
In one embodiment, there is provided a kind of method, the method include：Reception includes overlapping hand-written sequence of pen Draw, and for each stroke in this multiple stroke, by processor, the geometrical property based on the serial stroke determine with The associated plurality of feature of current stroke.The method of the embodiment is also based on the feature being associated with stroke by a series of strokes It is divided into one or more groups of strokes.As mentioned above, every group of stroke is a part for character or character, but does not have one group of pen Picture includes the stroke from more than one character.
This also based on one or more groups of strokes that the serial stroke has been divided into can be by the method for one embodiment Row stroke is divided into one or more characters, so as to increased a series of efficiency that strokes are divided into character.According to exemplary Embodiment, the method can be determined and the associated plurality of spy of current stroke by being based only upon the geometrical property of the serial stroke Levy, so that it is determined that with the associated plurality of feature of current stroke.In one embodiment, the method can also cause to described group In at least some group display so that least one set is shown in the way of differing markedly from least another group.Therefore, weight Folded hand-written display image can be simplified to facilitate user and check for example newest stroke.
In one embodiment, the method will also be with the associated plurality of feature normalization of current stroke.In this respect, Can be based on a series of overall size of strokes by the plurality of feature normalization.It is determined that the plurality of feature being associated with current stroke When, the plurality of feature can be from the terminal by current stroke, the geometric center of current stroke, the starting point of next stroke, next stroke Geometric center, the minimum rectangle comprising current stroke, the minimum rectangle comprising next stroke and the minimum comprising front unicursal Select in the group that rectangle is constituted.
The method can be performed with increment mode or batch mode.In increment mode, the determination of multiple features and The step of segmentation of serial stroke, is repeated after each continuous stroke is received.In batch mode, multiple features It is determined that repeating after multiple strokes are received the step of a series of segmentation with strokes.
In another embodiment, there is provided a kind of device, including at least one processor and including computer program generation At least one memorizer of code.At least one memorizer of the device of the embodiment and computer program code are configured to： Together with least one processor, the device is at least received including a series of Chong Die hand-written strokes, and it is multiple for this Each stroke in stroke, the geometrical property based on the serial stroke determine and the associated plurality of feature of current stroke.Should At least one memorizer of the device of embodiment and computer program code are additionally configured to together with least one processor So that a series of strokes are divided into one or more groups of strokes based on the feature being associated with stroke by the device.As above mentioned , every group of stroke is a part for character or character, but does not have one group of stroke to include the stroke from more than one character.
At least one memorizer of the device of one embodiment and computer program code are also configured to：With at least One processor causes one or more groups of strokes that the device has been divided into based on the serial stroke together by the serial stroke One or more characters are divided into, so as to increased a series of efficiency that strokes are divided into character.One exemplary embodiment party At least one memorizer of the device of formula and computer program code are additionally configured to：This is caused together with least one processor Device by be based only upon the serial stroke geometrical property determine with the associated plurality of feature of current stroke, it is determined that with current pen Draw associated plurality of feature.In one embodiment, at least one memorizer and computer program code of the device be also Can be configured to：Together with least one processor, cause the device to cause the display organized at least some in described group, with So that least one set is shown in the way of differing markedly from least another group.Therefore, overlapping hand-written display image can be simple Change to facilitate user to check for example newest stroke.
In one embodiment, at least one memorizer and computer program code of the device is also configured to： Together with least one processor, cause the device will be with the associated plurality of feature normalization of current stroke.In this respect, may be used A series of overall size based on strokes is by multiple feature normalizations.It is it is determined that during feature associated plurality of with current stroke, many Individual feature can be from the terminal by current stroke, the geometric center of current stroke, the starting point of next stroke, the geometry of next stroke The heart, the minimum rectangle comprising current stroke, the minimum rectangle comprising next stroke and the minimum rectangle comprising front unicursal are constituted Group in select.
The analysis to stroke can be performed with increment mode or batch mode.In increment mode, multiple features are really A series of the step of segmentation of fixed and strokes, is repeated after each continuous stroke is received.It is in batch mode, multiple The step of determination of feature and a series of segmentation of strokes, is repeated after multiple strokes are received.
In yet another embodiment, there is provided a kind of equipment, which includes including that overlap is hand-written a series of for receiving The device of stroke, and for each stroke in this multiple stroke, the geometrical property based on the serial stroke determines and works as The device of the associated plurality of feature of front stroke.The equipment of the embodiment is may also include for based on the spy being associated with stroke Levy a series of device that strokes are divided into one or more groups of strokes.As mentioned above, every group of stroke is character or character A part, but do not have one group of stroke to include the stroke from more than one character.
The equipment of one embodiment may also include a series of one or more groups of pens for being divided into based on strokes The serial stroke is divided into picture the device of one or more characters, so as to increase a series of effect that strokes are divided into character Rate.In an illustrative embodiments, for determining that the device with the associated plurality of feature of current stroke is included for only Geometrical property based on the serial stroke determines the device with the associated plurality of feature of current stroke.In an embodiment In, the equipment may also include for cause in described group at least some group display device so that least one set quilt Shown in the way of differing markedly from least another group.Therefore, the hand-written display image of overlap can be simplified to facilitate user and look into See for example newest stroke.
In one embodiment, the equipment also include for by with the associated plurality of feature normalization of current stroke Device.In this respect, can be based on a series of overall size of strokes by multiple feature normalizations.It is determined that being associated with current stroke Multiple features when, multiple features can from the terminal by current stroke, the geometric center of current stroke, the starting point of next stroke, The geometric center of next stroke, the minimum rectangle comprising current stroke, the minimum rectangle comprising next stroke and include previous pen Select in the group that the minimum rectangle of picture is constituted.
The equipment can be with increment mode or batch mode analysis stroke.In increment mode, the determination of multiple features A series of the step of segmentation with strokes, is repeated after each continuous stroke is received.In batch mode, Duo Gete The step of determination for levying and a series of segmentation of strokes, is repeated after multiple strokes are received.
In yet, there is provided a kind of computer program, it include have be stored in calculating therein At least one computer-readable memory of machine executable code part.The computer-executable code part is included for receiving Including the code instructions for overlapping a series of hand-written strokes and for for each stroke in this multiple stroke, being based on The geometrical property of the serial stroke determines the code instructions with the associated plurality of feature of current stroke.The embodiment Computer-executable code part also include for based on the feature being associated with stroke by a series of strokes be divided into one group or The code instructions of multigroup stroke.As mentioned above, every group of stroke is a part for character or character, but does not have one group Stroke includes the stroke from more than one character.
The computer-executable code part of one embodiment is may also include for being divided based on a series of strokes Into one or more groups of strokes the serial stroke is divided into the code instructions of one or more characters, so as to increased by A series of strokes are divided into the efficiency of character.The computer-executable code part of one illustrative embodiments may also include use Determine and the associated plurality of feature of current stroke in the geometrical property by being based only upon the serial stroke, it is determined that and current stroke The code instructions of associated plurality of feature.In one embodiment, the computer-executable code part can also be wrapped Include for cause in described group at least some group display code instructions so that least one set is not by with substantially The mode for being same as at least another group shows.Therefore, the hand-written display image of overlap can be simplified to facilitate user and check for example most New stroke.
In one embodiment, the computer-executable code part is also included for will be associated with current stroke The code instructions of multiple feature normalizations.In this respect, can be based on a series of overall size of strokes by multiple characteristic standards Change.It is determined that during feature associated plurality of with current stroke, the plurality of feature can be from the terminal by current stroke, current stroke Geometric center, the starting point of next stroke, the geometric center of next stroke, the minimum rectangle comprising current stroke, comprising next Select in the group that the minimum rectangle of stroke and the minimum rectangle comprising front unicursal are constituted.
The computer program can be with increment mode or batch mode analysis stroke.It is in increment mode, multiple The step of determination of feature and a series of segmentation of strokes, is repeated after each continuous stroke is received.In batch mode In, the determination of multiple features and repeat after multiple strokes are received the step of the segmentation of the serial stroke.
Description of the drawings
Some illustrative embodiments of the disclosure have been described with generic term, now with reference to accompanying drawing, accompanying drawing is simultaneously It is not necessarily drawn to scale, wherein：
Fig. 1 is the explanation hand-written to the overlap of wherein four character overlaps；
Fig. 2 is the block diagram of equipment according to an illustrative embodiment of the invention；
Fig. 3 is the functional block diagram of the operation for performing according to an illustrative embodiment of the invention；
Fig. 4 is the flow chart for illustrating the operation for performing according to an illustrative embodiment of the invention；
Fig. 5 is the identified overlap of the feature to wherein three continuous strokes according to an illustrative embodiment of the invention Hand-written explanation；
Fig. 6 is that the overlap to Fig. 1 according to an illustrative embodiment of the invention is hand-written is processed to identification four The explanation of the mode of character；
Fig. 7 is the flow chart of the operation for illustrating that increment mode according to an illustrative embodiment of the invention is performed；And
Fig. 8 is the flow process of the operation for illustrating that the batch mode of another exemplary embodiment of the invention is performed Figure.
Some embodiments of the present invention are described more fully hereinafter with referring now to accompanying drawing, wherein, show this Some but not all embodiments of invention.In fact, the various embodiments of the present invention much multi-form can embody simultaneously And should not be construed as limited to each embodiment stated here；More precisely, thesing embodiments are provided so that this theory Bright book will meet the legal requirement being suitable for.From start to finish, same reference refers to same element.As it is used herein, Each embodiment of the invention, term " data ", " content ", " information " and similar term can be interchangeably used for referring to The data that generation can be transmitted, receives and/or be stored.Therefore, the use of any this term should not be taken to limit the present invention Each embodiment spirit and scope.
In addition, as it is used herein, term " circuit " is referred to：（a）The only circuit realiration of hardware（Such as, simulating electricity Road and/or digital circuit）；（b）Circuit and including the software that is stored on one or more computer-readable memories and/ Or firmware instructions（Respectively）The combination of computer program, one rise operate such that device perform it is described herein one or Multiple functions；And（c）Circuit, for example, a part for microprocessor or microprocessor, which needs the software or solid for operation Part, even if the software or firmware are not what is be physically present.This definition of " circuit " is made suitable for term whole here With including in any claim.As another example, as it is used herein, term " circuit " also include comprising one or The realization of multiple processors and/or its part and subsidiary software and/or firmware.As another example, as it is used herein, Term " circuit " also includes, for example, for the based band integrated circuit or application processor integrated circuit of mobile phone, or server, honeybee Similar integrated circuit in the nest network equipment, other network equipments and/or other computing devices.
As herein defined, refer to the permanent, storage medium of physics（For example, volatibility or non-volatile memory device）'s " computer-readable recording medium " can be different from " the computer-readable transmission medium " for referring to electromagnetic signal.
As shown in fig. 1, it is hand-written to be input into via user interface such as touch screen in an overlapping arrangement, in the overlap mode In, character is continuously overlapped, and without space or time interval between character.In this respect, the example of Fig. 1 is illustrated under the figure Listed in portion respectively and overlapped with the order of four characters of 100 signs.In this respect, show that 102 illustrate the defeated of the first character Enter, show the 104 overlap inputs for illustrating the first and second characters, show that 106 illustrate that the overlap of first, second, and third character is defeated Enter, and show the 108 overlap inputs for illustrating all four characters.
Overlap hand-written permission user and make full use of user interface, this for the finger using them for example relatively small and The user that handwriting input is provided on the low touch screen of resolution is particularly useful.In addition, overlapping hand-written using provides relatively natural With the quick technology for being used to receive handwriting input.With reference to the gradually crowded display of Fig. 1, it will be recognized that due to other word Symbol and character overlap before, thus depend on overlap hand-written and chosen for the segmentation of kinds of characters is proposed with providing handwriting input War.Further, since showing while a lot of other overlap characters, overlap is hand-written also to cause user to check the current book of user The character or previous character write becomes extremely difficult.Therefore, it is used for there is provided one kind according to an illustrative embodiment of the invention Pre-segmentation overlaps hand-written technology.
Can by pre-segmentation overlap it is hand-written can receive via any one in various input equipments, such as user connects Mouthful, such as touch screen etc..Not only overlap it is hand-written can receive via various different input equipments, and these input equipments Can be proved and may include a part for various types of electronic equipment.For example, Fig. 2 illustrates the example that can embody the present invention The block diagram of the mobile terminal 10 of property embodiment.It will be appreciated, however, that the mobile terminal 10 for illustrating and being described below is only Only it is the example of a type of equipment that can benefit from the illustrative embodiments of the present invention, also, various types of movements Terminal, such as portable digital-assistant（PDA）, mobile phone, pager, mobile TV, game station, laptop computer, photographing unit, video recording Machine, audio/video player, radio, location equipment such as global positioning system（GPS）Equipment or above-mentioned and other types Sound and text communication system combination, can easily using the present invention illustrative embodiments.
Mobile terminal 10 may include the antenna 12 for operationally communicating with emitter 14 and receptor 16 or multiple antennas.Move Dynamic terminal 10 can further include to provide respectively signal extremely or receive the device of the spontaneous emitter 14 of signal and receptor 16, such as process Device 20.The signal include according to be suitable for cellular system air-interface standard signal message, and also including user speech, Receiving data and/or user generated data.In this respect, mobile terminal 10 can according to various first, second, third and/or Any one operation in forth generation communication protocol etc..For example, mobile terminal 10 can be according to the second filial generation（2G）Wireless communication protocol IS-136, time division multiple acess（TDMA）, global system for mobile communications（GSM）With IS-95 CDMAs（CDMA）, or according to the third generation （3G）Wireless communication protocol such as UMTS（UMTS）, CDMA2000, wideband CDMA（WCDMA）And time division synchronous CDMA（TD-CDMA）, according to 3.9G wireless communication protocols such as evolution UMTS Terrestrial Radio Access Networks（E-UTRAN）, according to Four generations（4G）Wireless communication protocol etc. is operated.Alternatively, or additionally, mobile terminal 10 can be according to non-cellular communication mechanisms Operation.For example, mobile terminal 10 can be in WLAN（WLAN）Or communicate in other communication networks.
In some embodiments, processor 20 may include expectation for realizing the audio frequency and logic function of mobile terminal 10 Circuit.For example, processor 20 includes one or more digital signal processors and/or one or more microprocessors.At this Reason device can further include one or more analog-digital converters, one or more digital to analog converters and/or other support circuits.Move The control of dynamic terminal 10 and signal processing function distribute between these devices according to the ability of each of which.Therefore, processor 20 may also include modulation and transmit before convolutional encoding and the staggeredly function of message and data.Processor 20 can comprise additionally in inside Speech coder, and may include internal data modem.Additionally, processor 20 may include that operation is storable in memorizer In one or more software programs function.For example, processor 20 is operable to connectivity program, such as traditional web browsing Device.Then, connectivity program can allow mobile terminal 10 according to such as WAP（WAP）, HTTP （HTTP）And/or similar agreement transmission and reception web content, such as location-based content and/or other web page contents.
Mobile terminal 10 may also include user interface, and the user interface includes outut device and input equipment, outut device For example, traditional earphone or speaker 24, ringer 22, mike 26, display 28, input equipment is, for example, user input Interface, these are all coupled to processor 20.The user input interface of permission 0 receiving data of mobile terminal 1 may include multiple Any one in the equipment of permission 0 receiving data of mobile terminal 1, such as keypad 30, touch-screen display（Such as with 28 table of display Show）, or other input equipments.In each embodiment including keypad 30, the keypad may include traditional numeral（0-9） Key and relative keys（#,*）And for operating other hardkeys and soft key of mobile terminal 10.Alternately, keypad 30 can also be wrapped Include various traditional qwerty keyboard arrangements.Keypad 30 may also include the soft key with correlation function.Additionally or alternatively, Mobile terminal 10 may include interface equipment, such as stick or other users input interface.Mobile terminal 10 can further include electricity Pond 34, such as vibrating battery group, for the various circuits for supplying electricity to operate mobile terminal 10 to need, and optionally for the machine of offer Tool vibration is used as detectable output.
As described above, the user input interface may include touch-screen display 28, which can be presented as any of touch Panel type display.Thus, for example, touch-screen display 28 may be configured as realizing touch recognition by any appropriate technology, it is such as electric The technologies such as resistance, electric capacity, infrared, deformeter, surface wave, optical imagery, decentralized signal technology, acoustic pulse recognition.Touch screen display Show that device 28 may be configured as the instruction of receiving user's input and the expression of user input be provided to processor 20.
Mobile terminal 10 can further include Subscriber Identity Module（UIM）38.UIM38 is typically with internal processor Storage device.UIM38 can for example include Subscriber Identity Module（SIM）, Universal Integrated Circuit Card（UICC）, common user mark Know module（USIM）, removable user identity module（R-UIM）Deng.UIM38 typically stores the information related to mobile subscriber Element.In addition to the UIM 38, mobile terminal 10 can also be equipped with memorizer.For example, mobile terminal 10 may include volatile storage Device 40, such as includes the volatile random access memory of the buffer zone of temporary transient storage for data（RAM）.Mobile terminal 10 Other nonvolatile memories 42 are may also include, which can be Embedded and/or can be removable.Memorizer can be stored The a plurality of information sum used by mobile terminal 10 performs each function of mobile terminal 10 according to this.For example, memorizer may include energy The identifier of mobile terminal 10, such as International Mobile Station Equipment Identification are uniquely identified enough（IMEI）Code.Additionally, these memorizeies can be deposited Store up for determining the instruction of cell id information.Specifically, these memorizeies can store the application journey for being performed by processor 20 Sequence, which determines the identity of the current area that mobile terminal 10 is in which in communicating, such as cell i d identity or cell id information.
The type of the type regardless of input equipment and the electronic equipment including the input equipment, overlapping handwriting input can It is analyzed according to illustrative embodiments as shown in Figure 3.In this embodiment, overlap is hand-written can be via touch screen 110 Receive, touch-screen display 28 of such as Fig. 2 etc..Then, each stroke of handwriting input can through feature extraction 112, wherein, such as It is described below to determine one or more features related to the stroke.Then, for example by grader 114 to current stroke it is whether With a part for front unicursal identical stroke group, or whether current stroke be that the part of another stroke group is determined.To the greatest extent Manage various graders 114 and can be used for analyzing the feature related to current stroke to determine the group belonging to current stroke, can adopt Conventional statistical sorter, such as support vector machine or artificial neural network.
As shown in Figure 3, grader 114 analyzes current stroke by using segmentation rule 116.Can determine in a variety of ways Segmentation rule 116.However, in the embodiment as shown, there is provided hand-written data storehouse 118, hand-written data storehouse 118 include multiple The handwritten stroke of different series.Then, each stroke series in hand-written data storehouse 118 can through feature extraction 120, wherein, point Each stroke series in the hand-written data storehouse is analysed to recognize the feature related to the stroke series and each including the series Stroke.In one embodiment, these features for being extracted and having associated with the current stroke in hand-written data storehouse 118 are through instruction Practice 122 to improve segmentation rule.In this respect, hand-written data storehouse 118 can comprising the various words by different person writings and/or The many samples of sentence.These words and/or sentence can be manually separated into character.Character is separated into based on this, each stroke Last stroke or the stroke in addition to last stroke of character of character can be classified as.Therefore, training 122 can Stroke learning from hand-written data storehouse 118 is defined on last stroke for determining stroke yes or no character to set up When parameter to be considered segmentation rule.It is as described below, the stroke can be represented by characteristic vector based on the various parameters of stroke, Then the value of this feature vector can be determined by grader 114 based on segmentation rule 116.Then, the comparable expression of the grader should The value of the characteristic vector of stroke and predefined threshold value are determining last stroke of the stroke yes or no character.Therefore, instruct Practice 122 to can be used to determine the maximum in terms of providing suitably specified stroke for the various strokes of whole in hand-written data storehouse 118 The parameter set of success rate.Based on training 122, segmentation rule 116 is defined, and which defines then and can be classified device 114 using with base The parameter set classified to stroke by the stroke of the different series stored in hand-written data storehouse 118, such as basis and each pen Draw the related feature of series and/or each stroke including each series.By comparing based on the spy for having extracted from current stroke The value of the characteristic vector levied and predefined threshold value, grader 114 can be it is thus determined that current stroke be and front unicursal identical pen A part or current stroke for picture group starts from different stroke groups.
Because purpose is that stroke is divided into complete character, if segmentation group only include character a part or If the group of person's segmentation includes the stroke from more than one character, error is likely to occur.Due to an exemplary embodiment party The method and apparatus of formula can be more easily adapted to conform to and correct any error related to the group of the segmentation for the part for only including character, Because the group of segmentation may be combined following pre-segmentation, in one embodiment, scalable predefines threshold value, for example By increasing the threshold value, to reduce the group of segmentation by the probability of the stroke included from more than one character.
In the illustrative embodiments of Fig. 3, the operation on dotted line includes the training stage 126.Thus, can first carry out in advance These operate and repeat these when need not be received by touch screen 110 handwriting input is overlapped and operate.However, under dotted line Operation include the execution stage 124 and overlap handwriting input reception when or perform afterwards, such as on touch screen 110 input After one or more strokes.
As further explanation, referring now to Fig. 4, there is shown with by dress according to an illustrative embodiment of the invention Put the operation of execution.The device for example can be employed on mobile terminal 10.Alternatively, however, the device may be embodied in respectively Plant in other equipment, including both mobile and fixing equipments, for example, any devices listed above.The device may include for Receive a series of device of strokes, such as processor 20, user input interface（Such as touch-screen display 28）Deng.Referring to Fig. 4's Operation 130.As described above, the multiple strokes for being received include that overlap is hand-written, wherein, multiple characters are connected one by one Continue out.For each stroke, it may be determined that multiple features.Referring to operation 132.In this respect, device may include for determine with The device of the related multiple features of each stroke, such as processor 20.Can determine that according to an illustrative embodiment of the invention various Feature.However, in one embodiment, for determined by current stroke, multiple features include the spy related to current stroke Levy and the feature related to previous and next stroke.In this respect, Fig. 5 is illustrated including current stroke 200, front unicursal 210 With multiple overlap hand-written characters of next stroke 220.
For example, for feature determined by current stroke may include the geometry of the terminal of current stroke, current stroke Center, the starting point of next stroke, the geometric center of next stroke, the minimum rectangle comprising current stroke, comprising next stroke Minimum rectangle and the minimum rectangle comprising front unicursal.Those features with regard to defining specified point or position, are generally sat with a pair Mark to define those features, such as x, y-coordinate.Similarly, for those features with rectangle or other two-dimensional shapes definition, each Shape such as each rectangle can be with four characterizing definitions, and such as the top of the coordinate and the shape in the left side and right side of the shape is with The coordinate in portion.With reference to Fig. 5, according to illustrative embodiments, the feature extracted from current stroke is may include with the current of 204 signs The x of the terminal of stroke, y-coordinate（endX,endY）, with the geometric center of the current stroke of 206 signs（currentGCX, currentGCY）, with the starting point of next stroke of 222 signs（startX,startY）, with the several of next stroke of 226 signs What center（nextGCX,nextGCY）With the minimum rectangle comprising current stroke, next stroke and front unicursal.Reference Fig. 5, often Individual rectangle can be by four characterizing definitions, i.e. the coordinate related to left side, right side, upper and lower.In this respect, in Figure 5, Left and right, the upper and lower coordinate of the minimum rectangle related to current stroke, next stroke and front unicursal is respectively with prefix current（Currently）、next（It is next）And previous（It is previous）Sign.For example, the minimum square with current stroke The related feature of shape is in Figure 5 with currentLeft（Work as front left）、currentRight（Work as front right）、currentTop（Currently On）And currentBottom（When front lower）Sign.Other points for being indicated to orient in Figure 5 include working as with 202 signs The starting point of front stroke, the terminal 224 of next stroke and respectively with 212 and 214 sign front unicursals beginning and end, although These other points are not extracted by as feature in the illustrative embodiments.
Although the geometric center of stroke can be defined in a variety of ways, according to an illustrative embodiments, stroke Geometric center be defined as stroke in the equalization point all put.For example, comprising terminal（xi, yi）（i=0…n-1）Pen Draw, with geometric center（GCX,GCY）, which is defined as foloows：
GCX=sum(xi)/n, GCY=sum (yi)/n；i=0…n-1
The character of multiple overlaps and the corresponding stroke including these characters may be of different sizes, because these characters May be written less or larger while still wanting to pass on identical implication.Thus, the device is may include for will be from current pen The feature normalization for extracting is drawn to consider（For example by removing）The device of different any impacts in identical stroke size, such as locates Reason device 20.Referring to the operation 134 of Fig. 4.In one embodiment, a series of overall size based on strokes is by these characteristic standards Change.In this respect, the minimum rectangle comprising each handwritten stroke can be defined as shown in Figure 5.In this respect, comprising whole pens The minimum rectangle of picture can be by the coordinate definition related to the up, down, left and right for overlapping hand-written character.Wherein comprising all series The minimum rectangle of stroke is positioned to a jiao at the origin with the coordinate system and with the side extended along the coordinate axess Illustrative embodiments in, the minimum rectangle comprising whole strokes can pass through its width such as totalWidth（Overall width）With Its height such as totalHeight（Total height）Represent in simplified form.
In one example, the feature having described above as follows can be standardized, wherein, prefix Current （Currently）、Next（It is next）And Pre（It is previous）Refer to the spy related to current stroke, next stroke and front unicursal respectively Levy.
Thus, CurrentStrokeEndX and CurrentStrokeEndY are the standardized x of the terminal of current stroke And y-coordinate.NextStrokeStartX and NextStrokeStartY are the standardized x and y coordinates of the starting point of next stroke. CurrentGCX and CurrentGCY are the standardized x and y coordinates of the geometric center of current stroke.NextGCX and NextGCY It is the standardized x and y coordinates of the geometric center of next stroke.CurrentLeft, CurrentRight, CurrentTop and CurrentBottom is left and right, the upper and lower coordinate of current stroke respectively.NextLeft, NextRight, NextTop and NextBottom is left and right, the upper and lower coordinate of next stroke respectively.Finally, PreLeft, PreRight, PreTop and PreBottom is left and right, the upper and lower coordinate of front unicursal respectively.
In one embodiment, then this multiple feature extracted from current stroke can be combined, such as by processing Device 20 is combined into characteristic vector.During just analyzed current stroke is the example of initial stroke wherein, previous pen is there will be no Draw.Thus, the feature related to front unicursal can be arranged to predefined value, such as -1.
Once for current stroke has determined that multiple features, and in one example, multiple strokes are standardized, then The serial stroke can be divided into by one or more groups of strokes based on the feature related to stroke.Referring to the operation 136 of Fig. 4.Thus, The device is may include for multiple serial strokes to be divided into the device of one or more groups of strokes, such as processor based on these features 20.As described above, these strokes can be divided into by grader 114 and according to the technology being described more fully below it is one or more groups of Stroke.The device may also include the one or more groups of strokes for being divided into based on multiple serial strokes, by multiple series Stroke is divided into the device of one or more characters, such as processor 20.Referring to operation 138.Segmentation based on these groups to character Can perform in a variety of ways, including by using segmentation rule 116 etc., 116 definition of segmentation rule will be classified device 114 and utilize Parameter suitably to estimate the characteristic vector of stroke.Thus, the processor 20 of an illustrative embodiments may include or Feature extraction 112, grader 114 and segmentation rule 116 are performed, as above in conjunction with Figure 3.Compared to not carrying out to stroke The example split by character is overlapped to multiple in the case of intermediate packets, by initially defining stroke group and being based on then this A little group definition characters, can perform stroke in the way of more having computational efficiency to the segmentation of character.
With reference to Fig. 6, multiple overlap characters are shown at 230.By multiple serial strokes are divided into stroke group, can be such as Multiple groups are defined shown in 232.Hereafter, multiple serial strokes can be divided into by multiple characters based on the group shown in 234.So Afterwards, the character of generation can be identified, such as by pattern recognition etc., as shown in 236, so that multiple overlap hand-written characters can be non- Chang Youxiao Ground Splits and it is identified as each series of characters.In the illustrative embodiments of Fig. 6, stroke is divided into based on these groups Character, can recognize the multiple possible combination of these groups, the such as such as possible combination of whole of these groups, 234 and 236 places Show.In this respect, can be recognized by handwriting recognition engine by each character that these various combinations organized are represented, the handwriting recognition engine Another computing device for communicating can be kept to embody by processor 20 or with the processor.In one embodiment, the hand-written knowledge Other engine may further determine that the similarity between every kind of combination of these groups and current character collection.Then, the combination table organized by these Each character set shown can be analyzed, and such as another computing device by communicating by processor 20 or with processor holding embodies Language model, to determine whether each character set meaningful.Then, the device（Such as processor 20）, handwriting recognition engine And/or language model, can be to based on the measurement of the similarity to having been determined by handwriting recognition engine, to being determined by language model The whether significant measurement of current character collection and in some embodiments to the measurement of the predefined geometrical property of each group Each possible character set of combination specifies fraction.In this respect, for example can analyze each group to determine by processor 20 Whether the group meets one or more predefined geometrical properties, and for example, a geometrical property can be related to the size of group, less than pre- The group for defining threshold value is considered as too little rather than complete character.As another example of geometrical property, positioned at along writing area The side in domain, such as along touch-screen display 28 most left or most right half group, complete character can be not considered as.For example it is based on One or more in similarity, meaningful and geometrical property are determined the character set with highest score by such as processor 20, can It is identified as most representing the plurality of character set for overlapping character, as indicated at 238.
In one embodiment, one or more limits can be applied to the multiple processes that may be combined for recognizing these groups System, to increase these combinations of identification and the efficiency with these combinations of post processing.For example, these possible combinations can be by appointing What potential combination or character are limited to maximum four groups and are limited.
Can overlap in addition to the efficiency split by character to multiple except increasing, sequence of pen signs in one or more groups of strokes Segmentation can also facilitate the display of handwriting input.In this respect, the device is may also include for causing at least some in these groups Show so that least one set is to differ markedly from the device shown in the way of in the of at least another group, such as processor 20, display 28 Deng.Referring to the operation 140 of Fig. 4.For example, the subset of only these groups can be shown, so that least one set is not displayed.At this Aspect, can show newest group of relatively small number purpose and whole groups before removing from display 28.Thus, display 28 is less Mixed and disorderly and user can more easily check the stroke and each stroke back to back before being currently input into.In other enforcements In mode, these stroke groups can be shown in the way of having significant difference between different stroke groups.For example, these stroke groups can use Different colours and/or luminance level show, the different colours or luminance level of the order change for such as being received based on stroke.One In individual illustrative embodiments, the available most dark color of newest stroke group（And/or brightness）Show, before back to back stroke The available slightly brighter color of group（And/or brightness）Show, the rest may be inferred until using most bright color（And/or brightness）Show Initial stroke group.Alternately, different stroke groups can be represented by different types of line, such as newest stroke indicated by the solid line Group, back to back stroke group before being represented with chain-dotted line, the like.In each embodiment, stroke group it is shown so that Different stroke groups are significantly different, and in some embodiments, newer stroke group is more obvious.
In one embodiment, the determination of the feature related to stroke and the stroke can be in inputs to the segmentation of current group Perform after each stroke, for example as shown in Figure 7.With regard to the subsequent descriptions of Fig. 7, current stroke and next stroke（As above Discuss）It is expressed as stroke k-1 and stroke k.With reference to the operation 250,252 and 254 of Fig. 7, after entering stroke k, it is determined that Whether the stroke is initial stroke, i.e. determine whether k=0.During the stroke being input into wherein is the example of initial stroke, can be just Beginningization actually writes region.Referring to operation 256.In this respect, actually write region to be initialized to surround the initial stroke Minimum rectangle, and the overall width and total height of the available rectangle for orienting relative to coordinate system as shown in Figure 5 and positioning or With the up, down, left and right coordinate definition of the rectangle positioned at other directions and position.Hereafter, enumerator k can wait next stroke Input before be incremented by.Referring to operation 258.After next stroke is input into, can recalculate and region actually be write so as to again What is calculated actually writes the minimum rectangle that region representation surrounds each stroke.Referring to operation 260.Hereafter, it may be determined that current stroke The feature of k-1.Referring to operation 262.Based on for the fixed features of stroke k-1, grader 114 can be based on the spy of stroke k-1 Offer value S is providedpre.In this respect, grader 114 can be based on the characteristic vector of stroke k-1 and regular 116 offer value S of segmentationpre, 116 definition of segmentation rule is by the grader using with the parameter of the characteristic vector of estimation stroke.Referring to the operation 264 of Fig. 7.So Afterwards, value S for being provided by grader 114preCan be with predefined threshold value T for classificationpreIt is compared.Referring to operation 266.If Value S provided by grader 114preMore than predefined threshold value T for classificationpre, then stroke k-1 and stroke k are considered to belong to Different groups.Referring to operation 268.If conversely, value S provided by grader 114preNo more than it is used for the predefined threshold of classification Value Tpre, then stroke k-1 and stroke k are considered to belong to identical group.Referring to operation 270.It is described above and figure 7 illustrates The process can incrementally repeat so that the serial stroke is suitably divided into group for each stroke of input.
As individually analyzing these strokes after the input of each stroke so as to suitably replacing these stroke groupings For scheme, can be with the multiple strokes of batch processing process analysis, such as shown in Figure 8.In this respect, it may be determined that actually write area Domain, such as surrounds multiple strokes 0,1 ..., the minimum rectangle of M-1.Referring to operation 280.Then, can initialization counter k, such as k=1, And can determine that whether whole strokes of this batch are considered, such as by comparing the number of the enumerator and stroke M.Referring to behaviour Make 282 and 284.In the example that whole strokes are also not accounted for wherein, determine the feature of stroke k-1 and can then be based on The feature of stroke k-1 for example determines value S of stroke k-1 by grader 114pre.Referring to operation 286 and 288.As before, can Carry out value SpreWith threshold value TpreComparison, wherein, stroke k-1 and stroke k belongs to different groups or identical group depends on respectively In value SpreWhether more than threshold value T for classificationpre.Referring to operation 290,292 and 294.Can be for 296 Counter k of operation Each stroke for being incremented by indicated this batch repeat the process, be taken into account and suitably divided up to each stroke Enter each group.
Fig. 4, Fig. 7 and Fig. 8 are the flow charts of methods and procedures product according to an illustrative embodiment of the invention.Will reason Solution, in each block and these flow charts of these flow charts, each piece of combination, can implement, for example firmly in a variety of ways Part, firmware, processor, circuit and/or related other of execution to the software including one or more computer program instructions Equipment.For example, one or more flow processs described above can be embodied by computer program instructions.In this respect, embody and retouch above The computer program instructions of each flow process stated, can be stored by the storage device of mobile terminal 10 and by the mobile terminal Reason device 20 is performed.It will be appreciated that any this computer program instructions can be loaded into computer or other programmable devices（Example Such as, hardware）On, to produce machine, so that these instruction generations performed on the computer or other programmable devices are used for Perform（Respectively）The device of specified function in flow chart block.These computer program instructions be also storable in bootable computer or In the computer-readable memory that other programmable devices are run in a specific way, to be stored in the computer-readable memory In these instruction produce include execution（Respectively）The product of the command device of specified function in flow chart block.These computer journeys Sequence instruction can also be loaded into computer or other programmable devices so that sequence of operations is in the computer or other are programmable Perform on device to produce computer implementation procedure, so as to these instructions performed on the computer or other programmable devices Specified function in execution flow chart block.
Therefore, the block of these flow charts supported for performing the combination of the device of specified function, specified for performing The operation of function and the combination for performing the programmed instruction of specified function.It will be further understood that one of these flow charts Or each piece of combination in multiple pieces, and these flow charts, can specified by performing function special hardware based computer System, or the combination of specialized hardware and computer instruction is implementing.
Under the beneficial teachings presented in description before and relevant drawings, the technical staff of field that the present invention belongs to will expect Many modifications of invention set forth herein and other embodiment.It will consequently be understood that, the present invention is not limited to disclosed Specific embodiment, and change and other embodiment be included within scope of the following claims.Although additionally, it Front description and relevant drawings describe each exemplary reality in the context of each element and/or some example combinations of function Apply mode, it should be appreciated that the various combination of each element and/or function can be also provided by alternate embodiments, without deviating from Scope of the following claims.In this respect, for example, each element and/or work(in addition to those for clearly describing above Can various combination it is contemplated and can be in the following claims some in propose.Although adopting specific art here Language, they only in a generic and descriptive sense using and be not used in restriction.
Priority Applications (1)
|Application Number||Priority Date||Filing Date||Title|
|PCT/CN2010/076285 WO2012024829A1 (en)||2010-08-24||2010-08-24||Method and apparatus for segmenting strokes of overlapped handwriting into one or more groups|
|Publication Number||Publication Date|
|CN103080878A CN103080878A (en)||2013-05-01|
|CN103080878B true CN103080878B (en)||2017-03-29|
Family Applications (1)
|Application Number||Title||Priority Date||Filing Date|
|CN201080068735.8A CN103080878B (en)||2010-08-24||2010-08-24||One or more groups of method and apparatus are divided into for hand-written stroke will be overlapped|
Country Status (4)
|JP (1)||JP5581448B2 (en)|
|KR (1)||KR101486174B1 (en)|
|CN (1)||CN103080878B (en)|
|WO (1)||WO2012024829A1 (en)|
Families Citing this family (10)
|Publication number||Priority date||Publication date||Assignee||Title|
|CN103376998B (en) *||2012-04-19||2016-06-15||中兴通讯股份有限公司||Handwriting equipment Chinese character type-setting method and device|
|JP5717691B2 (en) *||2012-05-28||2015-05-13||株式会社東芝||Handwritten character search device, method and program|
|JP2013246732A (en) *||2012-05-28||2013-12-09||Toshiba Corp||Handwritten character retrieval apparatus, method and program|
|EP2712433B1 (en) *||2012-05-31||2016-11-02||MultiTouch Oy||User interface for drawing with electronic devices|
|CN105283882B (en) *||2013-04-12||2019-12-27||诺基亚技术有限公司||Apparatus for text input and associated method|
|CN103345365B (en) *||2013-07-12||2016-04-13||北京蒙恬科技有限公司||The display packing of continuous handwriting input and the hand input device of employing the method|
|JP2015099566A (en) *||2013-11-20||2015-05-28||株式会社東芝||Feature calculation device, method and program|
|US9224038B2 (en)||2013-12-16||2015-12-29||Google Inc.||Partial overlap and delayed stroke input recognition|
|US9881224B2 (en)||2013-12-17||2018-01-30||Microsoft Technology Licensing, Llc||User interface for overlapping handwritten text input|
|CN105095924A (en) *||2014-04-25||2015-11-25||夏普株式会社||Handwriting recognition method and device|
Family Cites Families (17)
|Publication number||Priority date||Publication date||Assignee||Title|
|JPS60254384A (en) *||1984-05-31||1985-12-16||Fujitsu Ltd||Corresponding method of stroke|
|JPH05233599A (en) *||1992-02-18||1993-09-10||Seiko Epson Corp||Online character recognizing device|
|JPH06162266A (en) *||1992-11-20||1994-06-10||Seiko Epson Corp||Method for recognizing on-line handwritten character and device therefor|
|JPH0792817B2 (en) *||1994-08-19||1995-10-09||沖電気工業株式会社||Online character recognizer|
|JPH08161426A (en) *||1994-12-09||1996-06-21||Sharp Corp||Handwritten character stroke segmenting device|
|JP2939147B2 (en) *||1994-12-29||1999-08-25||シャープ株式会社||Handwritten character input display device and method|
|JPH09161011A (en) *||1995-12-13||1997-06-20||Matsushita Electric Ind Co Ltd||Handwritten character input device|
|US5970498A (en) *||1996-12-06||1999-10-19||International Business Machines Corporation||Object oriented framework mechanism for metering objects|
|JP3216800B2 (en) *||1997-08-22||2001-10-09||日立ソフトウエアエンジニアリング株式会社||Handwritten character recognition method|
|JP3024680B2 (en) *||1998-01-13||2000-03-21||日本電気株式会社||Handwritten pattern storage and retrieval device|
|JP3456931B2 (en) *||1999-12-10||2003-10-14||シャープ株式会社||Handwritten character recognition device, computer readable recording medium storing a handwritten character recognition program, and method for correcting characters recognized by handwritten characters|
|JP3974359B2 (en) *||2000-10-31||2007-09-12||株式会社東芝||Online character recognition apparatus and method, computer-readable storage medium, and online character recognition program|
|US7324691B2 (en) *||2003-09-24||2008-01-29||Microsoft Corporation||System and method for shape recognition of hand-drawn objects|
|JP2005141329A (en) *||2003-11-04||2005-06-02||Toshiba Corp||Device and method for recognizing handwritten character|
|KR100677426B1 (en) *||2005-01-14||2007-02-02||엘지전자 주식회사||Short message display method for mobile communication device|
|JP2009289188A (en)||2008-05-30||2009-12-10||Nec Corp||Character input device, character input method and character input program|
|CN101299236B (en) *||2008-06-25||2010-06-09||华南理工大学||Method for recognizing Chinese hand-written phrase|
Also Published As
|Publication number||Publication date|
|Goodfellow et al.||Multi-digit number recognition from street view imagery using deep convolutional neural networks|
|US9754184B2 (en)||Feature density object classification, systems and methods|
|US9274646B2 (en)||Method and apparatus for selecting text information|
|Zhang et al.||Touch gesture-based active user authentication using dictionaries|
|JP2017152015A (en)||Input assistance control system and method by slide-contact operation in portable terminal facility|
|US8204310B2 (en)||Feature design for HMM based Eastern Asian character recognition|
|Yang et al.||Improved deep convolutional neural network for online handwritten Chinese character recognition using domain-specific knowledge|
|Chou et al.||A binarization method with learning-built rules for document images produced by cameras|
|KR101264897B1 (en)||Character input method and system, electronic device and the keyboard thereof|
|US6721452B2 (en)||System and method of handwritten character recognition|
|CN100362525C (en)||Method for gathering and recording business card information in mobile phone by using image recognition|
|US7313277B2 (en)||Method and device for recognition of a handwritten pattern|
|US9135503B2 (en)||Fingertip tracking for touchless user interface|
|Arai et al.||Method for real time text extraction of digital manga comic|
|US8503788B2 (en)||Input-handwriting automatic transformation system and method|
|CN104166474B (en)||Information processor and character identifying method|
|Louloudis et al.||ICDAR 2011 writer identification contest|
|US7778464B2 (en)||Apparatus and method for searching for digital ink query|
|CN105283884A (en)||Classifying objects in digital images captured using mobile devices|
|Diem et al.||Icdar 2013 competition on handwritten digit recognition (hdrc 2013)|
|CN104809481B (en)||A kind of natural scene Method for text detection based on adaptive Color-based clustering|
|US7929770B2 (en)||Handwriting processing apparatus and method|
|CN103778250A (en)||Implement method for Chinese wubi cursive script dictionary query system|
|US9245191B2 (en)||System and method for scene text recognition|
|US20100238125A1 (en)||Method, Apparatus, and Computer Program Product For Discontinuous Shapewriting|
|C10||Entry into substantive examination|
|SE01||Entry into force of request for substantive examination|
|C41||Transfer of patent application or patent right or utility model|
|TA01||Transfer of patent application right||
Effective date of registration: 20160115
Address after: Espoo, Finland
Applicant after: Technology Co., Ltd. of Nokia
Address before: Espoo, Finland
Applicant before: Nokia Oyj