CN104915627B - A kind of character recognition method and device - Google Patents
A kind of character recognition method and device Download PDFInfo
- Publication number
- CN104915627B CN104915627B CN201410088106.1A CN201410088106A CN104915627B CN 104915627 B CN104915627 B CN 104915627B CN 201410088106 A CN201410088106 A CN 201410088106A CN 104915627 B CN104915627 B CN 104915627B
- Authority
- CN
- China
- Prior art keywords
- identified
- super
- text
- feature
- feature set
- 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.)
- Active
Links
Landscapes
- Character Discrimination (AREA)
Abstract
The embodiment of the invention discloses a kind of character recognition method and devices, which comprises receives any text to be identified of input；After the text to be identified is carried out Vector Processing, the feature set to be identified of the text to be identified is obtained, the feature set to be identified includes feature to be identified；The super side to be identified of predetermined number is generated according to the feature to be identified；Super super side while with pre-stored super in library to be identified is compared, when super super side while with described super in library to be identified, which matches number, meets preset condition, determines the Text region success to be identified.Compared with prior art, the present invention can be improved the recognition efficiency of text, while avoid the generation of identification error as far as possible.
Description
Technical field
The present invention relates to intelligent recognition fields, and in particular to a kind of character recognition method and device.
Background technique
Character recognition technology is text to be written by the handwriting input device of intelligent terminal, and extract character features letter
Breath, is finally compared with the characteristic information of pre-stored true text, to realize the technology for identifying the handwriting true and false.
Character recognition technology can be used for the identification of handwritten signature, and Handwritten Signature Recognition Method is a kind of authentication identification
Method, compared with other identity recognizing technologies, handwritten signature identification have contain much information, be not easy to imitate, accuracy is high and conveniently
Using the advantages that, so handwritten signature identification be it is a kind of by users approve safety identification authentication mode.
The Handwritten Signature Recognition Method of the prior art needs that actual signature is stored in advance, and passes through touch screen, hand in user
After writing plate or other handwriting input device input signature, handwritten signature identifying system can acquire the data information of signature, such as person's handwriting
Shape, writing speed write acceleration and writing pressure etc., then pre-process to the collected signed data information of institute,
Such as first stroke of a Chinese character processing merges, removes isolated point and redundant points, smooth and slant correction, to be eliminated as much as misleading recognition result
Factor, further, extracted from pretreated signed data information signature characteristic information and actual signature carry out
With comparison, to judge whether the signature of user meets authentication condition.
Although the identification of the signature inputted to user may be implemented in existing Handwritten Signature Recognition Method, still, due to
The recognition methods is more complicated to the Processing Algorithm of the signature of input, and committed memory is larger, needs after causing user to input signature
Want longer time that could complete identification work, recognition efficiency is lower.Meanwhile existing recognition methods accuracy of identification inaccuracy,
The probability that error occurs in identification is larger.
Summary of the invention
In view of the above-mentioned problems, can be improved the identification effect of text the present invention provides a kind of character recognition method and device
Rate, while the generation of identification error is avoided as far as possible.
The present invention provides a kind of character recognition methods, which comprises
Receive any text to be identified of input；
The text to be identified is subjected to Vector Processing；
Determine that the feature set to be identified of the text to be identified, the feature set to be identified include feature to be identified；
The super side to be identified of predetermined number is generated according to the feature to be identified；
Super super side while with pre-stored super in library to be identified is compared, when the super side to be identified with
When matching number when super super in library and meeting preset condition, the Text region success to be identified is determined.
Preferably, the method also includes:
Learn and store in advance the super Bian Ku of text, it is described it is super while library include the predetermined number it is super while.
Preferably, the preparatory super Bian Ku learnt and store text, the super side library include the super of the predetermined number
Side, comprising:
Obtain the same text that any user repeatedly inputs；
After the text is carried out Vector Processing, the feature set of the text is obtained, the feature set includes the text
Feature；
The super side of the predetermined number is generated according to the feature of the text, and forms super Bian Ku using the super side generated.
Preferably, the feature according to the text generates the super side of the predetermined number, and utilizes the super side generated
Form super Bian Ku, comprising:
The super side that multiple stochastical sampling generates the predetermined number is carried out to the feature of the text；
Super-network classifier is substituted using super side, after the super side is carried out learning classification, obtains super Bian Ku, the super side
Library includes the super side after learning classification.
Preferably, the super side to be identified that predetermined number is generated according to the feature to be identified, comprising:
Super-network classifier is substituted using super side, the super side to be identified of predetermined number is generated according to the feature to be identified.
Preferably, any text to be identified for receiving input, comprising:
Any text to be identified of input is received using stroke tracer technique.
Preferably, described by after the text progress Vector Processing to be identified, obtain the to be identified of the text to be identified
Feature set, the feature set to be identified include feature to be identified, comprising:
According to stroke, the text to be identified is divided into the characteristic segments of preset quantity；
Using least-square fitting approach, the direction vector of the characteristic segments is determined；
The feature set to be identified of the text to be identified, the spy to be identified are determined according to the direction vector of the characteristic segments
Collection includes feature to be identified.
Preferably, after the progress Vector Processing by the text, the feature of the text is obtained, comprising:
According to stroke, the text is divided into the characteristic segments of preset quantity；
Using least-square fitting approach, the direction vector of the characteristic segments is determined；
The feature set of the text is determined according to the direction vector of the characteristic segments.
The present invention also provides a kind of character recognition device, described device includes:
Receiving module, any text to be identified for receiving input；
Processing module, for the text to be identified to be carried out Vector Processing；
Determining module, for determining that the feature set to be identified of the text to be identified, the feature set to be identified include institute
State feature to be identified；
Generation module, for generating the super side to be identified of predetermined number according to the feature to be identified；
Contrast module works as institute for comparing super super side while with pre-stored super in library to be identified
When stating super super side while with described super in library to be identified and matching number and meet preset condition, the Text region to be identified is determined
Success.
Preferably, described device further include:
Study module, for learning and storing the super Bian Ku of text in advance, the super side library includes the predetermined number
Super side.
Preferably, the study module includes:
Acquisition submodule, the same text repeatedly inputted for obtaining any user；
Submodule is handled, after the text is carried out Vector Processing, obtains the feature set of the text, the feature
Collection includes the feature of the text；
First generates submodule, for generating the super side of predetermined number according to the feature of the text, and utilizes generation
Super side forms super Bian Ku.
Preferably, the first generation submodule includes:
Submodule is sampled, multiple stochastical sampling is carried out for the feature to the text and generates the super of the predetermined number
Side；
First classification submodule, for substituting super-network classifier using super side, after the super side is carried out learning classification,
Obtain super Bian Ku, it is described super in super after library includes learning classification.
Preferably, the generation module includes:
Second generates submodule, for substituting super-network classifier using super side, is generated according to the feature to be identified pre-
If the super side to be identified of number.
Preferably, the receiving module receives any to be identified of user's input particularly for using stroke tracer technique
The module of text.
Preferably, the processing module includes:
First divides submodule, for according to stroke, the text to be identified to be divided into the characteristic segments of preset quantity；
First determines submodule, for utilizing least-square fitting approach, determines the direction vector of the characteristic segments；
Second determines submodule, for determining the to be identified of the text to be identified according to the direction vector of the characteristic segments
Feature set, the feature set to be identified include feature to be identified.
Preferably, the processing submodule includes:
Second divides submodule, for according to stroke, the text to be divided into the characteristic segments of preset quantity；
Third determines submodule, for utilizing least-square fitting approach, determines the direction vector of the characteristic segments；
4th determines submodule, for determining the feature set of the text according to the direction vector of the characteristic segments.
The present invention learns in advance and stores the super Bian Ku of text, it is described it is super while library include predetermined number it is super while；It receives and uses
Any text to be identified of family input；After the text to be identified is carried out vector characteristics processing, the text to be identified is obtained
Vector characteristics collection to be identified, the vector characteristics collection to be identified includes the feature of the text to be identified；According to described wait know
The feature that other vector characteristics are concentrated generates super side to be identified；Super super side while with described super in library to be identified is carried out pair
Than determining described to be identified when super super side while with described super in library to be identified, which matches number, meets preset condition
Text region success.Compared with prior art, the present invention can be improved the recognition efficiency of text, while avoid identification error as far as possible
Generation.
Detailed description of the invention
In order to more clearly explain the technical solutions in the embodiments of the present application, make required in being described below to embodiment
Attached drawing is briefly described, it should be apparent that, the drawings in the following description are only some examples of the present application, for
For those of ordinary skill in the art, without any creative labor, it can also be obtained according to these attached drawings
His attached drawing.
Fig. 1 is the character recognition method flow chart that the embodiment of the present invention one provides；
Fig. 2 is the in-plane vectogram that the embodiment of the present invention one provides；
Fig. 3 is character recognition method flow chart provided by Embodiment 2 of the present invention；
Fig. 4 is input interface schematic diagram provided by Embodiment 2 of the present invention；
Fig. 5 is the character recognition device structure chart that the embodiment of the present invention three provides；
Fig. 6 is the block diagram of the part-structure for the relevant mobile phone of terminal that the embodiment of the present invention three provides.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present application, technical solutions in the embodiments of the present application carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of embodiments of the present application, instead of all the embodiments.It is based on
Embodiment in the application, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall in the protection scope of this application.
Embodiment one
With reference to Fig. 1, Fig. 1 is character recognition method flow chart provided in this embodiment, is specifically included:
S101: any text to be identified of input is received.
In the present embodiment, user can be used computer, smart phone or tablet computer input any one or it is more
A text is as text to be identified.
Specifically, the present embodiment can use stroke tracer technique receive user input text to be identified, to it is described to
It identifies that the false lift pen in text carries out auto-complete, tracks, passed through when lift pen is recorded all since starting to write a little
Point.
S102: the text to be identified is subjected to Vector Processing；
S103: determine that the feature set to be identified of the text to be identified, the feature set to be identified include spy to be identified
Sign.
In the present embodiment, after the text to be identified for receiving user's input, the text to be identified is carried out at vector
Reason, to obtain the feature set to be identified of the text to be identified.
In practical operation, the present embodiment is when receiving the text to be identified, using stroke tracer technique to described wait know
False lift pen in other text carries out auto-complete, tracks since starting to write a little, all points passed through when lift pen is recorded,
Then a series of this orderly point is equally divided into N sections, and every slope over 10 is calculated using least square fitting according to every section of point
Value, so that it is determined that direction vector out, finally obtains the feature set of the text to be identified.
Specifically, the text to be identified is divided into the characteristic segments of preset quantity first, in accordance with stroke by the present embodiment；Its
It is secondary, using least-square fitting approach, determine the direction vector of the characteristic segments；Finally, according to the direction of the characteristic segments to
Amount determines the feature set of the text to be identified.
Specifically, plane space is divided into 8 regions by the present embodiment, one direction of each Regional Representative can construct 8
A direction vector, is illustrated in fig. 2 shown below, and Fig. 2 is in-plane vectogram.Using least-square fitting approach, each feature is calculated
The corresponding slope of section, judges the slope is in which region of Fig. 2, according to affiliated area determine the directions of the characteristic segments to
Amount.
S104: the super side to be identified of predetermined number is generated according to the feature to be identified.
In the present embodiment, after the feature to be identified for obtaining the text to be identified, generated according to the feature to be identified
Super side to be identified.Specifically, substituting super-network classifier using super side, super side to be identified is generated according to the feature to be identified.
Wherein, super side can be through the generation of super-network classifier random initializtion.
In practical operation, using it is super while substitute super-network classifier sort it is super while in adaptive value be lower than give certain
The super side of a threshold value regenerates, and classifies again to all super sides, until meet all super sides it is reliable or reach to
Until fixed the number of iterations.
S105: super super side while with pre-stored super in library to be identified is compared, when described to be identified
When super super side while with described super in library matches number and meets preset condition, the Text region success to be identified is determined.
In the present embodiment, learn and store the super Bian Ku of text in advance, it is described it is super while library include predetermined number it is super while.
Due to character recognition technology be the text to be identified and pre-stored true text that will be inputted characteristic information into
Row compares, to realize the text true and false for identifying input.So several true texts of the preparatory learning user input of the present embodiment
Word, and the result of study is trained to obtain the super Bian Ku being made of super side, it is finally that the true text of user's input is corresponding
Super side library storage.
In practical application, for any one user, multiple input of the available user to the same text.Tool
Body, the number of input with no restrictions, generally higher than 10 times.The text that the user is repeatedly inputted carries out at vector characteristics
Reason, obtains the feature of the text.The super side of predetermined number is generated according to the feature of the text, and utilizes the super side group generated
At super Bian Ku.
Wherein, the present embodiment can the feature first to the text carry out multiple stochastical sampling and generate the super of predetermined number
Side；Secondly, substituting super-network classifier using super side, after the super side is carried out learning classification, super Bian Ku, the super side are obtained
Library includes the super side after learning classification.
In the present embodiment, the same text and one have predetermined number it is super while it is super while inventory in corresponding relationship.
Super-network classifier is a kind of cognitive learning model based on hypergraph model, and evolution super-network is initially as one kind
Parallel associative memory model is suggested, and is calculated and realized by DNA.The model can store by largely super Bian Zucheng, super side
The partial information of training set data expresses the correlation degree between the feature of sample and sample class, therefore is highly suitable for solving
The certainly pattern recognition problem of high dimensional data.Super-network has been successfully applied to solve various pattern recognition problems, such as text at present
Classification, the classification of handwriting digital optical identification data set etc..The learning method of super-network classifier includes super side method of substitution, is used
The method of random search searches for super side, will not cover this class with the super side of other type in the case where super-network is initial bad
Super side, but regenerate new super side at random from this classification again so that evolutional learning process may search for it is bigger
Space ensure that each classification has and make contributions to super side library.The present embodiment substitutes super-network classifier using super side, by institute
After the feature progress learning classification for stating text, feature after being classified；It is raw that multiple stochastical sampling is carried out to feature after the classification
Super Bian Ku is formed at the super side of predetermined number, and using the super side generated.
Compared with prior art, the present embodiment using super side substitute super-network classifier can handwriting characteristic to text into
Row learning classification, searching for, there is the best super side of decision-making capability to form super Bian Ku.
In the present embodiment, behind the super side to be identified for obtaining the text to be identified, by the super side to be identified and in advance
It is compared when super super in library of storage.The successful condition of Text region to be identified is preset, comparing result is worked as
When meeting preset condition, the Text region success to be identified is determined.
Specifically, can be by condition setting are as follows: when there are 1/3 super super sides while with described super in library to be identified
When successful match, the Text region success to be identified.
The present embodiment learns in advance and stores the super Bian Ku of text, it is described it is super while library include predetermined number it is super while；It receives
Any text to be identified of user's input；After the text to be identified is carried out vector characteristics processing, the text to be identified is obtained
The vector characteristics collection to be identified of word, the vector characteristics collection to be identified include the feature of the text to be identified；According to it is described to
Identify that the feature that vector characteristics are concentrated generates super side to be identified；By super super Bian Jinhang while with described super in library to be identified
Comparison determines described wait know when super super side while with described super in library to be identified, which matches number, meets preset condition
Other Text region success.Compared with prior art, the present embodiment can be improved the recognition efficiency of text, while avoid identifying as far as possible
The generation of error.
Embodiment two
With reference to Fig. 3, Fig. 3 is character recognition method flow chart provided in this embodiment, and the method is applied to signature identification
Field specifically includes:
S301: user inputs the signature of oneself in interface, and study may be selected or identify.
In the present embodiment, use Android system as user's interaction platform, and customize input interface as needed.User can
To input oneself signature, such as Fig. 4 in the input interface of the customization, user can be pressed by clicking the study below interface
The learning manipulation of button triggering signature.
S302: if user selects study, Vector Processing, the feature set signed are carried out to the signature of user.
When user selects to learn, Vector Processing, the feature set signed are carried out to the signature of the user.Specifically
, it is tracked since the starting to write a little of user's signature, all points passed through when lift pen is recorded are then a series of orderly this
Point be equally divided into N sections, and every slope over 10 value is calculated using least square fitting according to every section of point, so that it is determined that direction out
Vector finally obtains characteristic set.
S303: after user inputs multiple signatures, using super side replace super-network classifier to the feature set of signature into
Row training obtains super Bian Ku, described super when library includes super.
In order to enable more accurate when super super in library, the present embodiment requires user repeatedly to input signature, and to signature
Repeatedly learnt, obtains including the super Bian Ku with the best super side of decision-making capability until by training.
S304: if user selects identification, after the signature of user is carried out Vector Processing, obtain the signature wait know
Other feature set includes feature to be identified in the feature set to be identified.
When user selects recognition button after input signature, the signature of user is subjected to Vector Processing, obtains the label
The feature set to be identified of name.
The method that this step obtains feature set to be identified is the same as example 1, and details are not described herein.
S305: replacing super-network classifier using super side, is generated according to the feature set to be identified of the signature to be identified super
Side.
The present embodiment substitutes super-network classifier using super side, and the feature to be identified of the signature is carried out learning classification
Afterwards, the feature of the signature is obtained；Multiple stochastical sampling is carried out to the feature of the signature and generates the to be identified super of predetermined number
Side.
S306: matching when by super in library of the super Bian Yuchao to be identified, if successful match, the signature of user
Success identifies that otherwise nonrecognition is anyone signature.
Compared with prior art, the present embodiment can be improved the recognition efficiency of user's signature, while identification being avoided to miss as far as possible
The generation of difference.
Embodiment three
With reference to Fig. 5, Fig. 5 is character recognition device structure chart provided in this embodiment, and described device includes:
Receiving module 501, for receiving any text to be identified of user's input；
Processing module 502, for the text to be identified to be carried out Vector Processing；
Determining module 503, for determining that the feature set to be identified of the text to be identified, the feature set to be identified include
The feature to be identified；
Generation module 504, for generating super side to be identified according to the feature to be identified；
Contrast module 505, for super super side while with described super in library to be identified to be compared, when it is described to
When identifying that super super side while with described super in library matches number and meets preset condition, determine the Text region to be identified at
Function.
Wherein, the study module includes:
Acquisition submodule, the same text repeatedly inputted for obtaining any user；
Submodule is handled, after the text is carried out Vector Processing, obtains the feature set of the text, the feature
Collection includes the feature of the text；
First generates submodule, for generating the super side of predetermined number according to the feature of the text, and utilizes generation
Super side forms super Bian Ku.
Specifically, the first generation submodule includes:
Submodule is sampled, carries out the super side that multiple stochastical sampling generates predetermined number for the feature to the text；
First classification submodule, for substituting super-network classifier using super side, after the super side is carried out learning classification,
Obtain super Bian Ku, the super library Chao Bian in super after library includes learning classification, the super side library includes super after learning classification
Side.
The generation module includes:
Second classification submodule, for substituting super-network classifier using super side, by the spy in the feature set to be identified
After sign carries out learning classification, feature to be identified after being classified；
Second generates submodule, for generating super side to be identified according to feature to be identified after the classification.
In practical application, the receiving module particularly for using stroke tracer technique receive user input it is any to
Identify the module of text.
Specifically, the processing module includes:
First divides submodule, for according to stroke, the text to be identified to be divided into the characteristic segments of preset quantity；
First determines submodule, for utilizing least-square fitting approach, determines the direction vector of the characteristic segments；
Second determines submodule, for determining the to be identified of the text to be identified according to the direction vector of the characteristic segments
Feature set, the feature set to be identified include feature to be identified.
Specifically, the processing submodule includes:
Second divides submodule, for according to stroke, the text to be divided into the characteristic segments of preset quantity；
Third determines submodule, for utilizing least-square fitting approach, determines the direction vector of the characteristic segments；
4th determines submodule, for determining the feature set of the text according to the direction vector of the characteristic segments.
The embodiment of the invention also provides a kind of terminals, as shown in fig. 6, for ease of description, illustrating only and the present invention
The relevant part of embodiment, it is disclosed by specific technical details, please refer to present invention method part.The terminal can wrap
Include mobile phone, tablet computer, PDA(Persona lDigita lAssistant, personal digital assistant), POS
Any terminal device such as (Point of Sales, point-of-sale terminal), vehicle-mounted computer, taking the terminal as an example:
Fig. 6 shows the block diagram of the part-structure of mobile phone relevant to terminal provided in an embodiment of the present invention.With reference to figure
6, mobile phone includes: radio frequency (Radio Frequency, RF) circuit 610, memory 620, input unit 630, display list
First 640, sensor 650, voicefrequency circuit 660, Wireless Fidelity (wireless fidelity, WiFi) module 670, processor
The components such as 680 and power supply 690.It will be understood by those skilled in the art that handset structure shown in Fig. 6 does not constitute opponent
The restriction of machine may include perhaps combining certain components or different component layouts than illustrating more or fewer components.
It is specifically introduced below with reference to each component parts of the Fig. 6 to mobile phone:
RF circuit 610 can be used for receiving and sending messages or communication process in, signal sends and receivees, particularly, by base station
Downlink information receive after, to processor 680 handle；In addition, the data for designing uplink are sent to base station.In general, RF is electric
Road includes but is not limited to antenna, at least one amplifier, transceiver, coupler, low-noise amplifier (Low Noise
Amplifier, LNA), duplexer etc..In addition, RF circuit 610 can also by wireless communication with network and other set
Standby communication.Any communication standard or agreement, including but not limited to global system for mobile communications can be used in above-mentioned wireless communication
(G loba lSystem of Mobile communication, GSM), general packet radio service (Ge
Nera lPacket Radio Service, GPRS), CDMA (Code Division Multiple
Access, CDMA), wideband code division multiple access (Wideband Code Division Multiple Acces
S, WCDMA), long term evolution (Long Term Evolution, LTE)), Email, short message clothes
Be engaged in (Short Messaging Service, SMS) etc..
Memory 620 can be used for storing software program and module, and processor 680 is stored in memory 620 by operation
Software program and module, thereby executing the various function application and data processing of mobile phone.Memory 620 can mainly include
Storing program area and storage data area, wherein storing program area can application journey needed for storage program area, at least one function
Sequence (such as sound-playing function, image player function etc.) etc.；Storage data area can be stored to be created according to using for mobile phone
Data (such as audio data, phone directory etc.) etc..Storage data area in the present invention can be used for storing the super Bian Ku of text,
In, it is super while library include predetermined number it is super while.In addition, memory 620 may include high-speed random access memory, can also wrap
Include nonvolatile memory, a for example, at least disk memory, flush memory device or other volatile solid-state parts.
Input unit 630 can be used for receiving the number or character information of input, and generate the user setting with mobile phone 600
And the related key signals input of function control.Specifically, input unit 630 may include touch panel 631 and other inputs
Equipment 632.Touch panel 631, also referred to as touch screen, collecting the touch operation of user on it or nearby, (for example user makes
With the operation of any suitable object or attachment such as finger, stylus on touch panel 631 or near touch panel 631), and
Corresponding attachment device is driven according to preset formula.Touch panel 631 in the present invention can be used for receiving appointing for input
One text to be identified, user can be by inputting text to be identified by touch operation on it.Optionally, touch panel 631
It may include both touch detecting apparatus and touch controller.Wherein, the touch orientation of touch detecting apparatus detection user, and
Touch operation bring signal is detected, touch controller is transmitted a signal to；Touch controller is received from touch detecting apparatus
Touch information, and be converted into contact coordinate, then give processor 680, and order that processor 680 is sent can be received and added
To execute.Furthermore, it is possible to realize touch panel using multiple types such as resistance-type, condenser type, infrared ray and surface acoustic waves
631.In addition to touch panel 631, input unit 630 can also include other input equipments 632.Specifically, other input equipments
832 can include but is not limited to physical keyboard, function key (such as volume control button, switch key etc.), trace ball, mouse,
One of operating stick etc. is a variety of.
Display unit 640 can be used for showing information input by user or be supplied to user information and mobile phone it is various
Menu.Display unit 640 may include display panel 641, optionally, can use liquid crystal display (Liquid Crysta
LDisplay, LCD), Organic Light Emitting Diode (Organic Light-Emitting Diode, OLE
) etc. D forms configure display panel 641.Further, touch panel 631 can cover display panel 641, work as touch panel
After 631 detect touch operation on it or nearby, processor 680 is sent to determine the type of touch event, is then located
It manages device 680 and provides corresponding visual output on display panel 641 according to the type of touch event.Although in Fig. 6, touch surface
Plate 631 and display panel 641 are the input and input function for realizing mobile phone as two independent components, but in certain realities
Apply in example, can be integrated by touch panel 631 and display panel 641 and that realizes mobile phone output and input function.
Mobile phone 600 may also include at least one sensor 660, such as optical sensor, motion sensor and other sensings
Device.Specifically, optical sensor may include ambient light sensor and proximity sensor, wherein ambient light sensor can be according to environment
The light and shade of light adjusts the brightness of display panel 641, and proximity sensor can close display panel when mobile phone is moved in one's ear
641 and/or backlight.As a kind of motion sensor, accelerometer sensor can detect in all directions (generally three axis) and add
The size of speed can detect that size and the direction of gravity when static, can be used to identify application (such as the horizontal/vertical screen of mobile phone posture
Switching, dependent game, magnetometer pose calibrating), Vibration identification correlation function (such as pedometer, tap) etc.；Also as mobile phone
The other sensors such as configurable gyroscope, barometer, hygrometer, thermometer, infrared sensor, details are not described herein.
Voicefrequency circuit 660, loudspeaker 661, microphone 662 can provide the audio interface between user and mobile phone.Audio-frequency electric
Electric signal after the audio data received conversion can be transferred to loudspeaker 661, be converted to sound by loudspeaker 661 by road 660
Signal output；On the other hand, the voice signal of collection is converted to electric signal by microphone 662, is turned after being received by voicefrequency circuit 660
Audio data is changed to, then by after the processing of audio data output processor 680, through RF circuit 610 to be sent to such as the other hand
Machine, or audio data is exported to memory 620 to be further processed.
WiFi belongs to short range wireless transmission technology, and mobile phone can help user to receive and dispatch electricity by WiFi module 670
Sub- mail, browsing webpage and access streaming video etc., it provides wireless broadband internet access for user.Although Fig. 6 shows
Go out WiFi module 670, but it is understood that, and it is not belonging to must be configured into for mobile phone 600, it completely can be according to need
It to omit within the scope of not changing the essence of the invention.
Processor 680 is the control centre of mobile phone, using the various pieces of various interfaces and connection whole mobile phone, is led to
It crosses operation or executes the software program and/or module being stored in memory 620, and call and be stored in memory 620
Data execute the various functions and processing data of mobile phone, to carry out integral monitoring to mobile phone.Optionally, processor 680 can wrap
Include one or more processing units；Preferably, processor 680 can integrate application processor and modem processor, wherein answer
With the main processing operation system of processor, user interface and application program etc., modem processor mainly handles wireless communication.
It is understood that above-mentioned modem processor can not also be integrated into processor 680.
Mobile phone 600 further includes the power supply 690(such as battery powered to all parts), it is preferred that power supply can pass through electricity
Management system and processor 680 are logically contiguous, to realize management charging, electric discharge and power consumption by power-supply management system
The functions such as management.
Although being not shown, mobile phone 600 can also include camera, bluetooth module etc., and details are not described herein.
Specifically in the present embodiment, the processor 680 in terminal can be according to following instruction, will be one or more
The corresponding executable file of the process of application program is loaded into memory 620, and is run by processor 680 and be stored in storage
Application program in device 620, to realize various functions:
Receive any text to be identified of input；
After the text to be identified is carried out Vector Processing, the feature set to be identified of the text to be identified is obtained, it is described
Feature set to be identified includes feature to be identified；
The super side to be identified of predetermined number is generated according to the feature to be identified；
Super super side while with pre-stored super in library to be identified is compared, when the super side to be identified with
When matching number when super super in library and meeting preset condition, the Text region success to be identified is determined.
Preferably, the method also includes:
Learn and store in advance the super Bian Ku of text, it is described it is super while library include the predetermined number it is super while.
Preferably, the preparatory super Bian Ku learnt and store text, the super side library include the super of the predetermined number
Side, comprising:
Obtain the same text that any user repeatedly inputs；
After the text is carried out Vector Processing, the feature set of the text is obtained, the feature set includes the text
Feature,；
The super side of the predetermined number is generated according to the feature of the text, and forms super Bian Ku using the super side generated.
Preferably, the super side that predetermined number is generated according to the feature of the text, and utilize the super Bian Zucheng generated
Super Bian Ku, comprising:
The super side that multiple stochastical sampling generates the predetermined number is carried out to the feature of the text；
Super-network classifier is substituted using super side, after the super side is carried out learning classification, obtains super Bian Ku, the super side
Library includes the super side after learning classification.
Preferably, the super side to be identified that predetermined number is generated according to the feature to be identified, comprising:
Super-network classifier is substituted using super side, the super side to be identified of predetermined number is generated according to the feature to be identified.
Preferably, any text to be identified for receiving user's input, specifically:
Any text to be identified of user's input is received using stroke tracer technique.
Preferably, described by after the text progress Vector Processing to be identified, obtain the to be identified of the text to be identified
Feature set, the feature set to be identified include feature to be identified, comprising:
According to stroke, the text to be identified is divided into the characteristic segments of preset quantity；
Using least-square fitting approach, the direction vector of the characteristic segments is determined；
The feature set to be identified of the text to be identified, the spy to be identified are determined according to the direction vector of the characteristic segments
Collection includes feature to be identified.
Preferably, after the progress Vector Processing by the text, the feature of the text is obtained, comprising:
According to stroke, the text is divided into the characteristic segments of preset quantity；
Using least-square fitting approach, the direction vector of the characteristic segments is determined；
The feature set of the text is determined according to the direction vector of the characteristic segments.
Compared with prior art, character recognition device provided in this embodiment can be improved the recognition efficiency of text, simultaneously
The generation of identification error is avoided as far as possible.
For device embodiment, since it corresponds essentially to embodiment of the method, so related place is referring to method reality
Apply the part explanation of example.The apparatus embodiments described above are merely exemplary, wherein described be used as separation unit
The unit of explanation may or may not be physically separated, and component shown as a unit can be or can also be with
It is not physical unit, it can it is in one place, or may be distributed over multiple network units.It can be according to actual
It needs that some or all of the modules therein is selected to achieve the purpose of the solution of this embodiment.Those of ordinary skill in the art are not
In the case where making the creative labor, it can understand and implement.
It should be noted that, in this document, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to
Non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those
Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in process, method, article or equipment including the element.
It is provided for the embodiments of the invention character recognition method above and device is described in detail, it is used herein
A specific example illustrates the principle and implementation of the invention, and the above embodiments are only used to help understand
Method and its core concept of the invention；At the same time, for those skilled in the art is having according to the thought of the present invention
There will be changes in body embodiment and application range, in conclusion the content of the present specification should not be construed as to the present invention
Limitation.
Claims (14)
1. a kind of character recognition method, which is characterized in that the described method includes:
Receive any text to be identified of input；
The text to be identified is subjected to Vector Processing；
Determine that the feature set to be identified of the text to be identified, the feature set to be identified include feature to be identified；
The super side to be identified of predetermined number is generated according to the feature to be identified；
Super super side while with pre-stored super in library to be identified is compared, when the super side to be identified with it is described
When matching number when super super in library and meeting preset condition, the Text region success to be identified is determined；
The super side to be identified that predetermined number is generated according to the feature to be identified, comprising:
Super-network classifier is substituted using super side, after the feature to be identified in the feature set to be identified is carried out learning classification,
Feature to be identified after being classified, to feature to be identified after the classification carry out multiple stochastical sampling generate predetermined number wait know
Not super side.
2. the method according to claim 1, wherein the method also includes:
Learn and store in advance the super Bian Ku of text, it is described it is super while library include predetermined number it is super while.
3. according to the method described in claim 2, it is characterized in that, the preparatory super Bian Ku learnt and store text, described
It is super while library include predetermined number it is super while, comprising:
Obtain the same text that any user repeatedly inputs；
After the text is carried out Vector Processing, the feature set of the text is obtained, the feature set includes the spy of the text
Sign；
The super side of predetermined number is generated according to the feature of the text, and forms super Bian Ku using the super side generated.
4. according to the method described in claim 3, it is characterized in that, described generate predetermined number according to the feature of the text
Super side, and super Bian Ku is formed using the super side generated, comprising:
The super side that multiple stochastical sampling generates predetermined number is carried out to the feature of the text；
Super-network classifier is substituted using super side, after the super side is carried out learning classification, obtains super Bian Ku, the super Bian Kubao
Super side after including learning classification.
5. method according to any one of claims 1-4, which is characterized in that any text to be identified for receiving input
Word, comprising:
Any text to be identified of input is received using stroke tracer technique.
6. the method according to claim 1, wherein it is described the text to be identified is subjected to Vector Processing after,
The feature set to be identified of the text to be identified is obtained, the feature set to be identified includes feature to be identified, comprising:
According to stroke, the text to be identified is divided into the characteristic segments of preset quantity；
Using least-square fitting approach, the direction vector of the characteristic segments is determined；
The feature set to be identified of the text to be identified, the feature set to be identified are determined according to the direction vector of the characteristic segments
Including feature to be identified.
7. according to the method described in claim 3, it is characterized in that, obtaining institute after the progress Vector Processing by the text
State the feature set of text, comprising:
According to stroke, the text is divided into the characteristic segments of preset quantity；
Using least-square fitting approach, the direction vector of the characteristic segments is determined；
The feature set of the text is determined according to the direction vector of the characteristic segments.
8. a kind of character recognition device, which is characterized in that described device includes:
Receiving module, any text to be identified for receiving input；
Processing module, for the text to be identified to be carried out Vector Processing；
Determining module, for determining the feature set to be identified of the text to be identified, the feature set to be identified includes to be identified
Feature；
Generation module, for generating the super side to be identified of predetermined number according to the feature to be identified；
Contrast module, for super super side while with pre-stored super in library to be identified to be compared, when it is described to
When identifying that super super side while with described super in library matches number and meets preset condition, determine the Text region to be identified at
Function；
The generation module, be specifically used for using super side substitution super-network classifier, by the feature set to be identified wait know
After other feature carries out learning classification, feature to be identified after being classified is carried out feature to be identified after the classification repeatedly random
Sampling generates the super side to be identified of predetermined number.
9. device according to claim 8, which is characterized in that described device further include:
Study module, for learning and storing the super Bian Ku of text in advance, it is described it is super while library include predetermined number it is super while.
10. device according to claim 9, which is characterized in that the study module includes:
Acquisition submodule, the same text repeatedly inputted for obtaining any user；
Submodule is handled, after the text is carried out Vector Processing, obtains the feature set of the text, the feature set packet
Include the feature of the text；
First generates submodule, for generating the super side of predetermined number according to the feature of the text, and utilizes the super side generated
Form super Bian Ku.
11. device according to claim 10, which is characterized in that described first, which generates submodule, includes:
Submodule is sampled, carries out the super side that multiple stochastical sampling generates predetermined number for the feature to the text；
First classification submodule after the super side is carried out learning classification, is obtained for substituting super-network classifier using super side
Super Bian Ku is described super in super after library includes learning classification.
12. according to the device any in claim 8-11, which is characterized in that the receiving module is particularly for utilization
Stroke tracer technique receives the module of any text to be identified of user's input.
13. device according to claim 8, which is characterized in that the processing module includes:
First divides submodule, for according to stroke, the text to be identified to be divided into the characteristic segments of preset quantity；
First determines submodule, for utilizing least-square fitting approach, determines the direction vector of the characteristic segments；
Second determines submodule, for determining the feature to be identified of the text to be identified according to the direction vector of the characteristic segments
Collection, the feature set to be identified includes feature to be identified.
14. device according to claim 10, which is characterized in that the processing submodule includes:
Second divides submodule, for according to stroke, the text to be divided into the characteristic segments of preset quantity；
Third determines submodule, for utilizing least-square fitting approach, determines the direction vector of the characteristic segments；
4th determines submodule, for determining the feature set of the text according to the direction vector of the characteristic segments.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410088106.1A CN104915627B (en) | 2014-03-11 | 2014-03-11 | A kind of character recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410088106.1A CN104915627B (en) | 2014-03-11 | 2014-03-11 | A kind of character recognition method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104915627A CN104915627A (en) | 2015-09-16 |
| CN104915627B true CN104915627B (en) | 2019-04-26 |
Family
ID=54084685
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410088106.1A Active CN104915627B (en) | 2014-03-11 | 2014-03-11 | A kind of character recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104915627B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107391015B (en) * | 2017-07-19 | 2021-03-16 | 广州视源电子科技股份有限公司 | Control method, device and equipment of intelligent tablet and storage medium |
| CN108171144B (en) * | 2017-12-26 | 2020-12-11 | 四川大学 | Information processing method, information processing device, electronic equipment and storage medium |
| CN111079499B (en) * | 2019-07-02 | 2023-08-25 | 广东小天才科技有限公司 | Writing content identification method and system in learning environment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1619583A (en) * | 2003-11-20 | 2005-05-25 | 摩托罗拉公司 | Hand writing identifying method and system |
| CN102073870A (en) * | 2011-01-10 | 2011-05-25 | 杭州电子科技大学 | Method for recognizing Chinese character handwriting on touch screen |
| CN102298486A (en) * | 2010-06-22 | 2011-12-28 | 广东国笔科技股份有限公司 | Fast calling system and method based on touch screen |
| CN103136377A (en) * | 2013-03-26 | 2013-06-05 | 重庆邮电大学 | Chinese text classification method based on evolution super-network |
-
2014
- 2014-03-11 CN CN201410088106.1A patent/CN104915627B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1619583A (en) * | 2003-11-20 | 2005-05-25 | 摩托罗拉公司 | Hand writing identifying method and system |
| CN102298486A (en) * | 2010-06-22 | 2011-12-28 | 广东国笔科技股份有限公司 | Fast calling system and method based on touch screen |
| CN102073870A (en) * | 2011-01-10 | 2011-05-25 | 杭州电子科技大学 | Method for recognizing Chinese character handwriting on touch screen |
| CN103136377A (en) * | 2013-03-26 | 2013-06-05 | 重庆邮电大学 | Chinese text classification method based on evolution super-network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104915627A (en) | 2015-09-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10169639B2 (en) | Method for fingerprint template update and terminal device | |
| CN104239535B (en) | A kind of method, server, terminal and system for word figure | |
| CN103632165B (en) | A kind of method of image procossing, device and terminal device | |
| CN105912918B (en) | A kind of unlocked by fingerprint method and terminal | |
| CN104238918B (en) | List View component slippage display methods and device | |
| CN107077295A (en) | A kind of method, device, electronic equipment, display interface and the storage medium of quick split screen | |
| CN108875781A (en) | A kind of labeling method, apparatus, electronic equipment and storage medium | |
| CN104516887B (en) | A kind of web data searching method, device and system | |
| CN104217717A (en) | Language model constructing method and device | |
| CN107153818A (en) | Optical finger print verification method and Related product | |
| CN107729815A (en) | Image processing method, device, mobile terminal and computer-readable recording medium | |
| CN109426783A (en) | Gesture identification method and system based on augmented reality | |
| CN104123937A (en) | Method, device and system for reminding setting | |
| CN104217172B (en) | Privacy content inspection method and device | |
| CN107105093A (en) | Camera control method, device and terminal based on hand track | |
| CN107071129B (en) | A kind of bright screen control method and mobile terminal | |
| CN107590375A (en) | Unlocking method and related product | |
| CN107451450A (en) | Biometric discrimination method and Related product | |
| CN109543014B (en) | Man-machine conversation method, device, terminal and server | |
| CN104820546B (en) | Function information methods of exhibiting and device | |
| CN107516070A (en) | Biometric discrimination method and Related product | |
| CN109951889A (en) | A kind of Internet of Things matches network method and mobile terminal | |
| CN104915627B (en) | A kind of character recognition method and device | |
| CN108897482A (en) | A kind of information display method and terminal device | |
| CN105955507B (en) | A kind of display methods and terminal of soft keyboard |
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 |