CN104112130B - optical character recognition method and device - Google Patents
optical character recognition method and device Download PDFInfo
- Publication number
- CN104112130B CN104112130B CN201410299362.5A CN201410299362A CN104112130B CN 104112130 B CN104112130 B CN 104112130B CN 201410299362 A CN201410299362 A CN 201410299362A CN 104112130 B CN104112130 B CN 104112130B
- Authority
- CN
- China
- Prior art keywords
- stroke
- width
- value
- region
- point
- 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
Abstract
Present disclose provides a kind of optical character recognition method and device.Wherein, optical character recognition method includes：Character block is extracted from original image；Based on stroke width conversion, the stroke being broken in the character block is determined；Quasi- identification image is obtained to the stroke polishing of the fracture；Character recognition is carried out to the quasi- identification image, the text message in the quasi- identification image is obtained.The disclosure can be converted using stroke width and carry out polishing to the fracture that stroke in character occurs, and improve the recognition accuracy of optical character identification.
Description
Technical field
This disclosure relates to image processing field, in particular to optical character recognition method and device.
Background technology
OCR (Optical Character Recognition, optical character identification) technology is occupied in image processing field
Critical role, has a wide range of applications in information automation and office automation.OCR refers to text information, card etc.
Image file carries out analysis identifying processing, obtains the process of word and layout information.
It is critically important problem in OCR technique to improve recognition accuracy.In actual applications, it is necessary to carry out optical character knowledge
Often there is a situation where that stroke is broken in character in other original image, for example, in rechargeable card, user hangs to apply
When layer, residual is often had, causes the password of rechargeable card to be blocked, so as to cause to gesticulate fracture.The OCR of correlation technique holds
It is two strokes easily by the Stroke discrimination of fracture, causes recognition accuracy to reduce.
The content of the invention
The purpose of the disclosure is to provide optical character recognition method and device, the problem of to solve above-mentioned.
According to the embodiment of the present disclosure in a first aspect, present disclose provides a kind of optical character recognition method, including：
Character block is extracted from original image；
Based on stroke width conversion, the stroke being broken in the character block is determined；
Quasi- identification image is obtained to the stroke polishing of the fracture；
Character recognition is carried out to the quasi- identification image, the text message in the quasi- identification image is obtained.
The technical scheme provided by this disclosed embodiment, can be broken using stroke width conversion to stroke in character
Carry out polishing is split, the recognition accuracy of optical character identification is improved.
Above-mentioned optical character recognition method can also include following features, described based on stroke width conversion, it is determined that described
The stroke being broken in character block includes：
Breaking length threshold value is set；
The direction of stroke where determining the stroke area pixel point of the character block；
Judge that where the stroke area pixel point on the direction of stroke, the stroke area pixel point is next with it
It whether there is stroke width sudden change region between stroke area pixel point, and the length of the stroke width sudden change region is less than institute
Breaking length threshold value is stated, the stroke width sudden change region is white space, stroke width increment value more than the first width of setting
Change threshold or stroke width decreasing value are more than the region of the second change width threshold value of setting；
When there is the stroke width sudden change region, stroke where determining the stroke area pixel point is the pen of fracture
Draw, determine the region of fracture in the stroke that the stroke width sudden change region is the fracture.
The technical scheme provided by this disclosed embodiment, the stroke for determining fracture is converted using stroke width, can be accurate
Ground finds the region of fracture, and the stroke being broken for polishing provides basis, is favorably improved the recognition accuracy of character recognition.
Above-mentioned optical character recognition method can also include following features, and the stroke polishing to the fracture includes：
If direction of the region of fracture along stroke where it and the boundary pixel of stroke where it in the stroke of the fracture
Point be respectively the first point of interface and the second point of interface, by between first point of interface and second point of interface along described first
Stroke direction, the width of point of interface are equal to the stroke width value of first point of interface and the stroke width of second point of interface
The region of the average value of value sets the stroke region that is defined, wherein, the stroke width value of first point of interface and described second is handed over
The stroke width value of boundary's point becomes scaling method by stroke width and obtained；
The gray value in the quasi- stroke region is set to the value consistent with the gray value in stroke region in the character block.
The technical scheme provided by this disclosed embodiment, by by the gray value in quasi- stroke region be set to in character block
This mode of value that the gray value in stroke region is consistent carries out polishing to the stroke of fracture, can effectively eliminate fracture stroke, have
Help improve the recognition accuracy of character recognition.
Above-mentioned optical character recognition method can also include following features, and the stroke polishing to the fracture is also wrapped
Include：
When the region of fracture significantly increases region for the stroke width in the stroke of the fracture, by the fracture zone
The gray value in the stroke region in domain outside the quasi- stroke region is set to the ash with background area in the character block
The consistent value of angle value.
The technical scheme provided by this disclosed embodiment, can also further remove while stroke polishing is carried out and make an uproar
Sound, is favorably improved the recognition accuracy of character recognition.
Above-mentioned optical character recognition method can also include following features, and line character is entered to the quasi- identification image described
Before identification, in addition to：
Based on stroke width conversion, the noise in the character block is detected；
Eliminate the noise.
The technical scheme provided by this disclosed embodiment, can not only carry out polishing to the fracture that stroke in character occurs,
Noise can also be further removed using stroke width conversion, further increase the recognition accuracy of character recognition.
Above-mentioned optical character recognition method can also include following features, and described based on stroke width conversion, detection is described
Noise in character block includes：
The stroke width W of the character block is determined, character block i stroke width is wide equal to the first width average value and second
/ 2nd of average value sum are spent, the first width average value is character block i average stroke width value, character block i's
Average stroke width value is equal to the average value of the stroke width value of all stroke area pixel points in character block i, and described second is wide
Degree average value is the average value of the average stroke width value of all character blocks extracted from the original image, and i is nature
Number, the stroke width value of stroke area pixel point becomes scaling method by stroke width and obtained；
Character stroke width error d is set；
Judge stroke region that stroke width in the character block is in outside interval [W-d, W+d] whether meet it is following into
Make an uproar one of condition：The stroke that is not consistent in the setting range in the stroke region with the stroke direction, along the stroke
The stroke width change undulating value in region is more than setting fluctuation threshold or the stroke length in the stroke region is grown less than setting
Spend threshold value；
When the stroke region that stroke width is in outside interval [W-d, W+d] in the character block meets described into the condition of making an uproar
For the moment, judge that the stroke region that stroke width is in outside interval [W-d, W+d] in the character block is noise.
The technical scheme provided by this disclosed embodiment, the noise gone out using stroke width change detection in character block is
Eliminate noise and basis is provided, be favorably improved the recognition accuracy of character recognition.
Above-mentioned optical character recognition method can also include following features, and the elimination noise includes：
The gray value of the noise is set to the value consistent with the gray value of background area in the character block.
The technical scheme provided by this disclosed embodiment, can effectively eliminate the noise in character block, be favorably improved
The recognition accuracy of character recognition.
Above-mentioned optical character recognition method can also include following features, and the stroke width, which becomes scaling method, to be included：
Rim detection is carried out to the character block, edge pixel point is found, the edge pixel point has a direction ladder
Angle value；
According to any edge pixel p gradient direction dpThe edge picture matched with edge pixel point p is searched along route r
Vegetarian refreshments q, edge pixel point q gradient direction dqWith edge pixel point p gradient direction dpOn the contrary, wherein, r=p+n*dp, n is big
In or equal to 0；
When finding the edge pixel point q, the edge pixel point p and edge pixel point q line is specified
The stroke width value of upper each pixel is the Euclidean distance between the edge pixel point p and edge pixel point q.
The technical scheme provided by this disclosed embodiment, can effectively determine stroke width value, be the stroke of polishing fracture
Basis is provided.
According to the second aspect of the embodiment of the present disclosure, present disclose provides a kind of optical character recognition device, including：
Extraction module, for extracting character block from original image；
Determining module, for being converted based on stroke width, determines the stroke being broken in the character block；
Polishing module, quasi- identification image is obtained for the stroke polishing to the fracture；
Identification module, for carrying out character recognition to the quasi- identification image, obtains the text in the quasi- identification image
Information.
The technical scheme provided by this disclosed embodiment, can be broken using stroke width conversion to stroke in character
Carry out polishing is split, the recognition accuracy of optical character identification is improved.
Above-mentioned optical character recognition device can also include following features, and the determining module includes：
Threshold setting unit, for setting breaking length threshold value；
Stroke direction determining unit, the direction of stroke where the stroke area pixel point for determining the character block；
It is mutated judging unit, the direction for judging the stroke where the stroke area pixel point, the stroke area
It whether there is stroke width sudden change region between domain pixel and its next stroke area pixel point, and the stroke width is prominent
The length for becoming region is less than the breaking length threshold value, and the stroke width sudden change region is white space, stroke width increase
Value is more than the first change width threshold value of setting or stroke width decreasing value is more than the region of the second change width threshold value of setting；
Stroke determining unit is broken, for when there is the stroke width sudden change region, determining stroke region picture
Stroke where vegetarian refreshments is the stroke of fracture, determines the fracture zone in the stroke that the stroke width sudden change region is the fracture
Domain.
The technical scheme provided by this disclosed embodiment, the stroke for determining fracture is converted using stroke width, can be accurate
Ground finds the region of fracture, and the stroke being broken for polishing provides basis, is favorably improved the recognition accuracy of character recognition.
Above-mentioned optical character recognition device can also include following features, and the polishing module includes：
Quasi- stroke area setting unit, if direction along stroke where it of the region of fracture in the stroke of the fracture and its
The boundary pixel of place stroke is respectively the first point of interface and the second point of interface, and the quasi- stroke area setting unit is used for will
It is equal to described first along stroke direction, the width of first point of interface between first point of interface and second point of interface
The region of the average value of the stroke width value of point of interface and the stroke width value of second point of interface sets the stroke region that is defined,
Wherein, the stroke width value of first point of interface and the stroke width value of second point of interface are converted single by stroke width
Member is obtained；
First gray scale adjusting unit, for the gray value in the quasi- stroke region to be set to and stroke in the character block
The consistent value of the gray value in region.
The technical scheme provided by this disclosed embodiment, by by the gray value in quasi- stroke region be set to in character block
This mode of value that the gray value in stroke region is consistent carries out polishing to the stroke of fracture, can effectively eliminate fracture stroke, have
Help improve the recognition accuracy of character recognition.
Above-mentioned optical character recognition device can also include following features, and the polishing module also includes：
Second gray scale adjusting unit, for significantly increasing for the stroke width in the stroke of the fracture when the region of fracture
During big region, by the gray value that the stroke region outside the quasi- stroke region is in the region of fracture be set to it is described
The consistent value of the gray value of background area in character block.
The technical scheme provided by this disclosed embodiment, can also further remove while stroke polishing is carried out and make an uproar
Sound, is favorably improved the recognition accuracy of character recognition.
Above-mentioned optical character recognition device can also include following features, before the identification module, in addition to：
Noise detection module, for being converted based on stroke width, detects the noise in the character block；
Cancellation module, for eliminating the noise that the noise detection module is detected.
The technical scheme provided by this disclosed embodiment, can not only carry out polishing to the fracture that stroke in character occurs,
Noise can also be further removed using stroke width conversion, further increase the recognition accuracy of character recognition.
Above-mentioned optical character recognition device can also include following features, and the noise detection module includes：
Stroke width determining unit, the stroke width W for determining the character block, character block i stroke width is equal to
/ 2nd of first width average value and the second width average value sum, the first width average value is the flat of character block i
Equal stroke width value, character block i average stroke width value is equal to the stroke width of all stroke area pixel points in character block i
The average value of value, the second width average value is that the average stroke of all character blocks extracted from the original image is wide
The average value of angle value, i is natural number, and the stroke width value of stroke area pixel point is obtained by stroke width converter unit；
Error setting unit, for setting character stroke width error d；
Into judging unit of making an uproar, for judging that stroke width is in the stroke outside interval [W-d, W+d] in the character block
Whether region meets following into one of condition of making an uproar：It is not consistent in the setting range in the stroke region with the stroke direction
Stroke, along the stroke region stroke width change undulating value be more than setting fluctuation threshold or the stroke region pen
Draw length and be less than setting length threshold；
Identifying unit, for expiring when the stroke region that stroke width is in outside interval [W-d, W+d] in the character block
Foot is described into the condition of making an uproar a period of time, judges that stroke width is in the stroke region outside interval [W-d, W+d] in the character block
For noise.
The technical scheme provided by this disclosed embodiment, the noise gone out using stroke width change detection in character block is
Eliminate noise and basis is provided, be favorably improved the recognition accuracy of character recognition.
Above-mentioned optical character recognition device can also include following features, and the cancellation module includes：
Gray scale is except unit of making an uproar, for the gray value of the noise to be set to the gray scale with background area in the character block
The consistent value of value.
The technical scheme provided by this disclosed embodiment, can effectively eliminate the noise in character block, be favorably improved
The recognition accuracy of character recognition.
Above-mentioned optical character recognition device can also include following features, and the stroke width converter unit includes：
Rim detection subelement, for carrying out rim detection to the character block, finds edge pixel point, the edge picture
Vegetarian refreshments has a direction gradient value；
Coupling subelement, for the gradient direction d according to any edge pixel ppSearched and edge pixel point along route r
The edge pixel point q that p matches, edge pixel point q gradient direction dqWith edge pixel point p gradient direction dpOn the contrary, its
In, r=p+n*dp, n is more than or equal to 0；
Subelement is specified, for when finding the edge pixel point q, specifying the edge pixel point p and the side
The stroke width value of each pixel is between the edge pixel point p and edge pixel point q on edge pixel q line
Euclidean distance.
The technical scheme provided by this disclosed embodiment, can effectively determine stroke width value, be the stroke of polishing fracture
Basis is provided.
According to the third aspect of the embodiment of the present disclosure there is provided a kind of optical character recognition device, including：
Processor；
Memory for storing processor-executable instruction；
Wherein, the processor is configured as performing and such as given an order：
Character block is extracted from original image；
Based on stroke width conversion, the stroke being broken in the character block is determined；
Quasi- identification image is obtained to the stroke polishing of the fracture；
Character recognition is carried out to the quasi- identification image, the text message in the quasi- identification image is obtained.
It should be appreciated that the general description of the above and detailed description hereinafter are only exemplary, this can not be limited
It is open.
Brief description of the drawings
Fig. 1 is a kind of one of flow chart of optical character recognition method according to an exemplary embodiment；
Fig. 2 is the two of the flow chart of a kind of optical character recognition method according to an exemplary embodiment；
Fig. 3 is a kind of one of structured flowchart of optical character recognition device according to an exemplary embodiment；
Fig. 4 is a kind of knot of determining module 20 in a kind of optical character recognition device according to an exemplary embodiment
Structure block diagram；
Fig. 5 is the structural frames of polishing module 30 in a kind of optical character recognition device according to an exemplary embodiment
One of figure；
Fig. 6 is the structural frames of polishing module 30 in a kind of optical character recognition device according to an exemplary embodiment
The two of figure；
Fig. 7 is the structural frames of extraction module 10 in a kind of optical character recognition device according to an exemplary embodiment
Figure；
Fig. 8 is the two of the structured flowchart of a kind of optical character recognition device according to an exemplary embodiment；
Fig. 9 is the knot of noise detection module 50 in a kind of optical character recognition device according to an exemplary embodiment
Structure block diagram；
Figure 10 is the structure of cancellation module 60 in a kind of optical character recognition device according to an exemplary embodiment
Block diagram；
Figure 11 is stroke width converter unit in a kind of optical character recognition device according to an exemplary embodiment
70 structured flowchart；
Figure 12 is a kind of block diagram of device according to an exemplary embodiment.
Accompanying drawing herein is merged in specification and constitutes the part of this specification, shows the implementation for meeting the present invention
Example, and for explaining principle of the invention together with specification.
Embodiment
The disclosure is described in further detail below by specific embodiment and with reference to accompanying drawing.
Fig. 1 is a kind of one of flow chart of optical character recognition method according to an exemplary embodiment.Such as Fig. 1 institutes
Show, in the present embodiment, the flow of optical character recognition method may include steps of：
In step S101, character block is extracted from original image；
The character block extracted includes stroke region and background area., can be by the pixel in stroke region in the disclosure
Gray value be set to 0, the gray value of background area is set to 255, the gray value of the pixel in stroke region can also be set to
255, the gray value of background area is set to 0.
Wherein, step S101 can be realized using following sub-step (step a11 to step d11)：
A11, gray processing processing post filtering is carried out to original image, the first pre-processed results figure of original image is obtained；
B11, binary conversion treatment post filtering is carried out to the first pre-processed results figure, obtains the second pretreatment knot of original image
Fruit is schemed；
Text message in c11, the second pre-processed results figure of positioning and extraction；
D11, is split by traveling line character to the text message in the second pre-processed results figure, obtains character block, the character
Block is the character block extracted from original image.
Step a11 to step d11 removes noise simultaneously during character block extraction is carried out, and is favorably improved character knowledge
Other recognition accuracy.
In step s 102, converted based on stroke width, determine the stroke being broken in the character block；
Wherein, step S102 can be realized using following sub-step (step a12 to step d12)：
In step a12, breaking length threshold value is set；
In step b12, the direction of stroke where determining the stroke area pixel point of character block；
In step c12, judge where stroke area pixel point on the direction of stroke, under stroke area pixel point and its
It whether there is stroke width sudden change region between one stroke area pixel point, and the length of the stroke width sudden change region is less than
Breaking length threshold value, wherein, stroke width sudden change region can be white space, stroke width significantly increases region or stroke is wide
Degree is substantially reduced region；Wherein, stroke width significantly increase region refer to stroke width increment value be more than setting the first width become
Change the region of threshold value, stroke width is substantially reduced region and refers to that stroke width decreasing value is more than the second change width threshold value of setting
Region.
In step d12, when there is stroke width sudden change region, stroke where determining stroke area pixel point is fracture
Stroke, determine stroke width sudden change region for fracture stroke in the region of fracture.
The stroke for determining fracture is converted using stroke width, the region of fracture can be correctly found, is the pen of polishing fracture
Draw and basis is provided, be favorably improved the recognition accuracy of character recognition.
In step s 103, quasi- identification image is obtained to the stroke polishing of fracture；
Wherein, step S103 can be realized using following sub-step (step a13 to step b13)：
In step a13, if direction of the region of fracture along stroke where it in the stroke of fracture and stroke where it
Boundary pixel is respectively the first point of interface and the second point of interface, will be had a common boundary between the first point of interface and the second point of interface along first
Stroke direction, the width of point are equal to the average value of the stroke width value of the first point of interface and the stroke width value of the second point of interface
Region sets the stroke region that is defined, wherein, the stroke width value of the first point of interface and the stroke width value of the second point of interface pass through
Stroke width becomes scaling method and obtained；
In step b13, the gray value in quasi- stroke region is set to consistent with the gray value in stroke region in character block
Value.
Step a13 to step b13 is by the way that the gray value in quasi- stroke region to be set to the ash with stroke region in character block
This mode of the consistent value of angle value carries out polishing to the stroke of fracture, can effectively eliminate fracture stroke, be favorably improved character
The recognition accuracy of identification.
In the disclosed embodiments, step S103 can also include sub-step c13：
In step c13, when the stroke width in stroke of the region of fracture for fracture significantly increases region, by fracture zone
The gray value in the stroke region in domain outside quasi- stroke region is set to consistent with the gray value of background area in character block
Value.So, noise can also be further removed while stroke polishing is carried out, the identification for being favorably improved character recognition is accurate
True rate.
In step S104, alignment identification image carries out character recognition, obtains the text message in quasi- identification image.
Here, illustratively stroke width becomes scaling method.Stroke width becomes scaling method and may include steps of (step
A00 to step c00)：
In step a00, rim detection is carried out to character block, edge pixel point is found, edge pixel point has a side
To Grad；
In step b00, according to any edge pixel p gradient direction dpSearched and edge pixel point p phases along route r
The edge pixel point q of matching, edge pixel point q gradient direction dqWith edge pixel point p gradient direction dpOn the contrary, wherein, r
=p+n*dp, n is more than or equal to 0；
In step c00, when finding edge pixel point q, designated edge pixel p and edge pixel point q line
The stroke width value of upper each pixel is the Euclidean distance between edge pixel point p and edge pixel point q.Edge pixel point p
Euclidean distance between edge pixel point q is expressed as | | p-q | |.
The stroke width, which becomes scaling method, can effectively determine stroke width value, and the stroke being broken for polishing provides basis.
Optical character recognition method shown in the present embodiment, can convert what is occurred to stroke in character using stroke width
Fracture carries out polishing, improves the recognition accuracy of optical character identification.
Fig. 2 is the two of the flow chart of a kind of optical character recognition method according to an exemplary embodiment.Such as Fig. 2 institutes
Show, in the present embodiment, the flow of optical character recognition method may include steps of：
In step s 201, character block is extracted from original image；
In step S202, based on stroke width conversion, the stroke being broken in character block is determined；
In step S203, based on stroke width conversion, the noise in character block is detected；
Wherein, step S203 can be realized using following sub-step (step a23 to step d23)：
In step a23, the stroke width W of character block is determined, character block i stroke width is equal to the first width average value
With 1/2nd of the second width average value sum, wherein, the first width average value is character block i average stroke width value,
Character block i average stroke width value is equal to the average value of the stroke width value of all stroke area pixel points in character block i, the
Two width average values are the average value of the average stroke width value of all character blocks extracted from original image, and i is nature
Number, the stroke width value of stroke area pixel point becomes scaling method by stroke width and obtained；
If character block i stroke width is Wi, the first width average value is Wi1, the second width average value is Wa, then, Wa
Expression formula such as formula (1) shown in：
Formula (1)
In formula (1), i is natural number, and n is the character block sum extracted from original image.
WiExpression formula such as formula (2) shown in：
Formula (2)
In step b23, character stroke width error d is set；
In step c23, judge whether the stroke region that stroke width is in outside interval [W-d, W+d] in character block is full
Foot is following into one of condition of making an uproar：The stroke that is not consistent in the setting range in stroke region with the stroke direction, along the stroke
The stroke width change undulating value in region is more than setting fluctuation threshold or the stroke length in the stroke region is less than setting length
Threshold value；
In step d23, when the stroke region that stroke width is in outside interval [W-d, W+d] in character block meets above-mentioned
Into the condition of making an uproar for the moment, judge that the stroke region that stroke width is in outside interval [W-d, W+d] in the character block is noise.
The noise that step a23 is gone out in character block to step d23 using stroke width change detection, base is provided to eliminate noise
Plinth, is favorably improved the recognition accuracy of character recognition.
In step S204, the noise detected in step S203 is eliminated；
In the disclosed embodiments, the noise detected in step S203 can be eliminated in the following way：By noise
Gray value be set to the value consistent with the gray value of background area in character block.This mode can effectively eliminate character block
In noise, be favorably improved the recognition accuracy of character recognition.
In step S205, quasi- identification image is obtained to the stroke polishing of fracture；
In step S206, alignment identification image carries out character recognition, obtains the text message in quasi- identification image.
In the disclosed embodiments, step S203 and step S204 position are not fixed, after may be at step S201,
Before step S206, after step S203 and step S204 can also be placed on to step S205, before step S206.
Optical character recognition method shown in the present embodiment, can not only be mended to the fracture that stroke in character occurs
Together, additionally it is possible to further remove noise using stroke width conversion, further increase the recognition accuracy of character recognition.
One of ordinary skill in the art will appreciate that all or part of processing in above method embodiment can be by
The related hardware of programmed instruction is completed, and foregoing program can be stored in a kind of computer read/write memory medium, the program
Upon execution, the step of including preceding method embodiment is performed, and foregoing storage medium includes：ROM, RAM, magnetic disc or CD
Etc. it is various can be with the medium of store program codes.
Fig. 3 is a kind of one of structured flowchart of optical character recognition device according to an exemplary embodiment.Such as Fig. 3
Shown, in the present embodiment, optical character recognition device can include extraction module 10, determining module 20, polishing module 30 and know
Other module 40.Wherein, extraction module 10 is used to extract character block from original image.Determining module 20 is used to be based on stroke width
Conversion, determines the stroke being broken in the character block that extraction module 10 is extracted.Polishing module 30 is used for what determining module 20 was determined
The stroke polishing of fracture obtains quasi- identification image.The quasi- identification image that identification module 40 is used to obtain the polishing of polishing module 30 enters
Line character is recognized, obtains the text message in quasi- identification image.
Optical character recognition device shown in the present embodiment, can convert what is occurred to stroke in character using stroke width
Fracture carries out polishing, improves the recognition accuracy of character recognition.
Fig. 4 is a kind of knot of determining module 20 in a kind of optical character recognition device according to an exemplary embodiment
Structure block diagram.As shown in figure 4, in the present embodiment, determining module 20 can include threshold setting unit 210, stroke direction and determine list
Member 220, mutation judging unit 230 and fracture stroke determining unit 240.Wherein, threshold setting unit 210 is used to set breaking length
Spend threshold value.Stroke direction determining unit 220 is used for the direction of stroke where determining the stroke area pixel point of character block.Mutation is sentenced
Disconnected unit 230 is used for the direction for judging the stroke where stroke area pixel point, stroke area pixel point and its next pen
It whether there is stroke width sudden change region between picture area pixel point, and the length of the stroke width sudden change region is set less than threshold value
The breaking length threshold value of the setting of unit 210 is put, the stroke width sudden change region can be big for white space, stroke width increment value
It is more than the region of the second change width threshold value of setting in the first change width threshold value of setting or stroke width decreasing value.It is broken stroke
Determining unit 240 is used for when there is stroke width sudden change region, and stroke where determining stroke area pixel point is the pen of fracture
Draw, it is the region of fracture in the stroke of fracture to determine the stroke width sudden change region.
The stroke for determining fracture is converted using stroke width, the region of fracture can be correctly found, is the pen of polishing fracture
Draw and basis is provided, be favorably improved the recognition accuracy of character recognition.
Fig. 5 is the structural frames of polishing module 30 in a kind of optical character recognition device according to an exemplary embodiment
One of figure.As shown in figure 5, in the present embodiment, polishing module 30 can include the quasi- gray scale of stroke area setting unit 310 and first
Adjustment unit 320.Quasi- stroke area setting unit 310 includes stroke width converter unit 70.If disconnected in the stroke of fracture
The boundary pixel of direction of the split plot domain along stroke where it and stroke where it is respectively the first point of interface and the second point of interface,
Quasi- stroke area setting unit 310 be used for by between the first point of interface and second point of interface along the stroke side of the first point of interface
The region for being equal to the average value of the stroke width value of the first point of interface and the stroke width value of the second point of interface to, width is set to
Quasi- stroke region, wherein, the stroke width value of the first point of interface and the stroke width value of the second point of interface are become by stroke width
Unit 70 is changed to obtain.First gray scale adjusting unit 320 is used to the gray value in quasi- stroke region being set to and stroke in character block
The consistent value of the gray value in region.
Embodiment illustrated in fig. 5 is by the way that the gray value in quasi- stroke region to be set to the gray value with stroke region in character block
This mode of consistent value carries out polishing to the stroke of fracture, can effectively eliminate fracture stroke, be favorably improved character recognition
Recognition accuracy.
Fig. 6 is the structural frames of polishing module 30 in a kind of optical character recognition device according to an exemplary embodiment
The two of figure.As shown in fig. 6, in the present embodiment, polishing module 30 can include quasi- stroke area setting unit 310, the first gray scale
The gray scale adjusting unit 330 of adjustment unit 320 and second.Quasi- stroke area setting unit 310, first gray scale adjusting unit 320
Function is identical with Fig. 5, and here is omitted.Second gray scale adjusting unit 330 is used for when in stroke of the region of fracture for fracture
Stroke width when significantly increasing region, the gray value in the stroke region being in the region of fracture outside quasi- stroke region is set
For the value consistent with the gray value of background area in character block.So, can also further it be gone while stroke polishing is carried out
Except noise, the recognition accuracy of character recognition is favorably improved.
Fig. 7 is the structural frames of extraction module 10 in a kind of optical character recognition device according to an exemplary embodiment
Figure.As shown in fig. 7, in the present embodiment, gray processing unit 110, binarization unit 120, positioning can be included in extraction module 10
With extraction unit 130, Character segmentation unit 140.Gray processing unit 110 is used to original image is carried out to filter after gray processing processing
Ripple, obtains the first pre-processed results figure of original image.Binarization unit 120 is used to carry out two-value to the first pre-processed results figure
Change processing post filtering, obtain the second pre-processed results figure of original image.Positioning and extraction unit 130 are used to positioning and extracting the
Text message in two pre-processed results figures.Character segmentation unit 140 is used to split text message by traveling line character, obtains
Character block, the character block is the character block that extraction module 10 is extracted.
Optical character recognition device shown in the present embodiment, removes noise simultaneously during character block extraction is carried out,
It is favorably improved the recognition accuracy of character recognition.
Fig. 8 is the two of the structured flowchart of a kind of optical character recognition device according to an exemplary embodiment.Such as Fig. 8
Shown, in the present embodiment, optical character recognition device can include extraction module 10, determining module 20, polishing module 30, identification
Module 40, noise detection module 50 and cancellation module 60.Wherein, extraction module 10, determining module 20, polishing module 30 and identification
Module 40 it is identical with Fig. 3, here is omitted.Noise detection module 50 is used to, based on stroke width conversion, detect character
Noise in block.Cancellation module 60 is used to eliminate the noise that noise detection module 50 is detected.
Optical character recognition device shown in the present embodiment, can not only be mended to the fracture that stroke in character occurs
Together, additionally it is possible to further remove noise using stroke width conversion, further increase the recognition accuracy of character recognition.
Fig. 9 is the knot of noise detection module 50 in a kind of optical character recognition device according to an exemplary embodiment
Structure block diagram.As shown in figure 9, in the present embodiment, noise detection module 50 can include stroke width determining unit 510, error and set
Put unit 520, into judging unit 530 and the identifying unit 540 of making an uproar.Stroke width determining unit 510 includes stroke width conversion
Unit 70.Wherein, stroke width determining unit 510 is used for the stroke width W for determining character block, character block i stroke width etc.
In 1/2nd of the first width average value and the second width average value sum, wherein, the first width average value is character block i's
Average stroke width value, character block i average stroke width value is wide equal to the stroke of all stroke area pixel points in character block i
The average value of angle value, the second width average value is the average stroke width value of all character blocks extracted from original image
Average value, i is natural number, and the stroke width value of stroke area pixel point is obtained by stroke width converter unit 70.Error is set
Putting unit 520 is used to set character stroke width error d.It is used to judge in character block at stroke width into judging unit 530 of making an uproar
Whether the stroke region outside interval [W-d, W+d] meets following into one of condition of making an uproar：Do not have in the setting range in stroke region
Have be consistent with stroke direction stroke, along the stroke region stroke width change undulating value be more than setting fluctuation threshold or
The stroke length in stroke region is less than setting length threshold.Identifying unit 540 is used to work as stroke width in character block and is in interval
Stroke region outside [W-d, W+d] meets above-mentioned into the condition of making an uproar a period of time, judges that stroke width is in interval in the character block
Stroke region outside [W-d, W+d] is noise.
The noise that embodiment illustrated in fig. 9 is gone out in character block using stroke width change detection, basis is provided to eliminate noise,
It is favorably improved the recognition accuracy of character recognition.
Figure 10 is the structure of cancellation module 60 in a kind of optical character recognition device according to an exemplary embodiment
Block diagram.As shown in Figure 10, in the present embodiment, cancellation module 60 can include gray scale except unit 610 of making an uproar.Gray scale is except unit 610 of making an uproar
For the gray value of noise to be set to the value consistent with the gray value of background area in character block.This mode can be effectively
The noise in character block is eliminated, the recognition accuracy of character recognition is favorably improved.
Figure 11 is stroke width converter unit in a kind of optical character recognition device according to an exemplary embodiment
70 structured flowchart.As shown in figure 11, in the present embodiment, stroke width converter unit 70 can include rim detection subelement
710th, coupling subelement 720 and specified subelement 730.Wherein, rim detection subelement 710 is used to carry out edge inspection to character block
Survey, find edge pixel point, edge pixel point has a direction gradient value.Coupling subelement 720 is used for according to any edge
Pixel p gradient direction dpThe edge pixel point q matched with edge pixel point p is searched along route r, edge pixel point q's
Gradient direction dqWith edge pixel point p gradient direction dpOn the contrary, wherein, r=p+n*dp, n is more than or equal to 0.Specify subelement
730 are used for when finding edge pixel point q, each pixel on designated edge pixel p and edge pixel point q line
Stroke width value is the Euclidean distance between edge pixel point p and edge pixel point q.Edge pixel point p and edge pixel point q it
Between Euclidean distance be expressed as | | p-q | |.
Embodiment illustrated in fig. 11 can effectively determine stroke width value, and the stroke being broken for polishing provides basis.
It will be appreciated by those skilled in the art that the module in device in embodiment can be distributed according to embodiment description
In the device of embodiment, respective change can also be carried out and be disposed other than in one or more devices of the present embodiment.Above-mentioned reality
The module for applying example can be merged into a module, can also be further split into multiple submodule.
Using the optical character recognition method and device of the disclosure, font that can be roughly equal to stroke width carries out light
Character recognition is learned, these fonts can be printing type face, rechargeable card, credit card font etc..
Figure 12 is a kind of block diagram of device according to an exemplary embodiment.For example, device 800 can be mobile electricity
Words, computer, digital broadcast terminal, messaging devices, game console, tablet device, Medical Devices, body-building equipment is individual
Personal digital assistant etc..
Reference picture 12, device 800 can include following one or more assemblies：Processing assembly 802, memory 804, power supply
Component 806, multimedia groupware 808, audio-frequency assembly 810, the interface 812 of input/output (I/O), sensor cluster 814, and
Communication component 816.
The integrated operation of the usual control device 800 of processing assembly 802, such as with display, call, data communication, phase
Machine operates the operation associated with record operation.Treatment element 802 can refer to including one or more processors 820 to perform
Order, to complete all or part of step of above-mentioned method, these steps include：Character block is extracted from original image；It is based on
Stroke width is converted, and determines the stroke being broken in the character block；Quasi- identification image is obtained to the stroke polishing of the fracture；It is right
The quasi- identification image carries out character recognition, obtains the text message in the quasi- identification image.In addition, processing assembly 802 can
Including one or more modules, to be easy to the interaction between processing assembly 802 and other assemblies.For example, processing component 802 can be with
Including multi-media module, to facilitate the interaction between multimedia groupware 808 and processing assembly 802.
Memory 804 is configured as storing various types of data supporting the operation in equipment 800.These data are shown
Example includes the instruction of any application program or method for being operated on device 800, and contact data, telephone book data disappears
Breath, picture, video etc..Memory 804 can be by any kind of volatibility or non-volatile memory device or their group
Close and realize, such as static RAM (SRAM), Electrically Erasable Read Only Memory (EEPROM) is erasable to compile
Journey read-only storage (EPROM), programmable read only memory (PROM), read-only storage (ROM), magnetic memory, flash
Device, disk or CD.
Electric power assembly 806 provides electric power for the various assemblies of device 800.Electric power assembly 806 can include power management system
System, one or more power supplys, and other components associated with generating, managing and distributing electric power for device 800.
Multimedia groupware 808 is included in the screen of one output interface of offer between described device 800 and user.One
In a little embodiments, screen can include liquid crystal display (LCD) and touch panel (TP).If screen includes touch panel, screen
Curtain may be implemented as touch-screen, to receive the input signal from user.Touch panel includes one or more touch sensings
Device is with the gesture on sensing touch, slip and touch panel.The touch sensor can not only sensing touch or sliding action
Border, but also detection touches or slide related duration and pressure with described.In certain embodiments, many matchmakers
Body component 808 includes a front camera and/or rear camera.When equipment 800 be in operator scheme, such as screening-mode or
During video mode, front camera and/or rear camera can receive the multi-medium data of outside.Each front camera and
Rear camera can be a fixed optical lens system or with focusing and optical zoom capabilities.
Audio-frequency assembly 810 is configured as output and/or input audio signal.For example, audio-frequency assembly 810 includes a Mike
Wind (MIC), when device 800 be in operator scheme, when such as call model, logging mode and speech recognition mode, microphone by with
It is set to reception external audio signal.The audio signal received can be further stored in memory 804 or via communication set
Part 816 is sent.In certain embodiments, audio-frequency assembly 810 also includes a loudspeaker, for exports audio signal.
I/O interfaces 812 is provide interface between processing assembly 802 and peripheral interface module, above-mentioned peripheral interface module can
To be keyboard, click wheel, button etc..These buttons may include but be not limited to：Home button, volume button, start button and lock
Determine button.
Sensor cluster 814 includes one or more sensors, and the state for providing various aspects for device 800 is commented
Estimate.For example, sensor cluster 814 can detect opening/closed mode of equipment 800, the relative positioning of component is for example described
Component is the display and keypad of device 800, and sensor cluster 814 can be with 800 1 components of detection means 800 or device
Position change, the existence or non-existence that user contacts with device 800, the orientation of device 800 or acceleration/deceleration and device 800
Temperature change.Sensor cluster 814 can include proximity transducer, be configured to detect in not any physical contact
The presence of neighbouring object.Sensor cluster 814 can also include optical sensor, such as CMOS or ccd image sensor, for into
As being used in application.In certain embodiments, the sensor cluster 814 can also include acceleration transducer, gyro sensors
Device, Magnetic Sensor, pressure sensor or temperature sensor.
Communication component 816 is configured to facilitate the communication of wired or wireless way between device 800 and other equipment.Device
800 can access the wireless network based on communication standard, such as WiFi, 2G or 3G, or combinations thereof.In an exemplary implementation
In example, communication component 816 receives broadcast singal or broadcast related information from external broadcasting management system via broadcast channel.
In one exemplary embodiment, the communication component 816 also includes near-field communication (NFC) module, to promote junction service.Example
Such as, NFC module can be based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra wide band (UWB) technology,
Bluetooth (BT) technology and other technologies are realized.
In the exemplary embodiment, device 800 can be believed by one or more application specific integrated circuits (ASIC), numeral
Number processor (DSP), digital signal processing appts (DSPD), PLD (PLD), field programmable gate array
(FPGA), controller, microcontroller, microprocessor or other electronic components are realized, for performing the above method.
In the exemplary embodiment, a kind of non-transitorycomputer readable storage medium including instructing, example are additionally provided
Such as include the memory 804 of instruction, above-mentioned instruction can be performed to complete the above method by the processor 820 of device 800.For example,
The non-transitorycomputer readable storage medium can be ROM, random access memory (RAM), CD-ROM, tape, floppy disk
With optical data storage devices etc..
Embodiments of the invention additionally provide a kind of non-transitorycomputer readable storage medium, when in the storage medium
Instruction by the computing device of mobile terminal when so that mobile terminal is able to carry out a kind of date storage method, the data
Storage method includes：Character block is extracted from original image；Based on stroke width conversion, the pen being broken in the character block is determined
Draw；Quasi- identification image is obtained to the stroke polishing of the fracture；Character recognition is carried out to the quasi- identification image, the standard is obtained
Recognize the text message in image.
Preferred embodiment of the present disclosure is the foregoing is only, the disclosure is not limited to, for the skill of this area
For art personnel, the disclosure can have various modifications and variations.It is all within the spirit and principle of the disclosure, that is made any repaiies
Change, equivalent substitution, improvement etc., should be included within the protection domain of the disclosure.
Claims (15)
1. a kind of optical character recognition method, it is characterised in that including：
Character block is extracted from original image；
Based on stroke width conversion, the stroke being broken in the character block is determined；
If direction of the region of fracture along stroke where it and the boundary pixel point of stroke where it in the stroke of the fracture
Not Wei the first point of interface and the second point of interface, will between first point of interface and second point of interface along described first have a common boundary
Stroke direction, the width of point are equal to the stroke width value of first point of interface and the stroke width value of second point of interface
The region of average value sets the stroke region that is defined, wherein, the stroke width value of first point of interface and second point of interface
Stroke width value pass through stroke width become scaling method obtain；
The gray value in the quasi- stroke region is set to the value consistent with the gray value in stroke region in the character block to obtain
Image is recognized to standard；
Character recognition is carried out to the quasi- identification image, the text message in the quasi- identification image is obtained.
2. optical character recognition method according to claim 1, it is characterised in that described to be converted based on stroke width, really
The stroke being broken in the fixed character block includes：
Breaking length threshold value is set；
The direction of stroke where determining the stroke area pixel point of the character block；
Judge where the stroke area pixel point on the direction of stroke, the stroke area pixel point and its next stroke
It whether there is stroke width sudden change region between area pixel point, and the length of the stroke width sudden change region is less than described break
Length threshold is split, the stroke width sudden change region is white space, stroke width increment value more than the first change width of setting
Threshold value or stroke width decreasing value are more than the region of the second change width threshold value of setting；
When there is the stroke width sudden change region, stroke where determining the stroke area pixel point is the stroke of fracture,
Determine the region of fracture in the stroke that the stroke width sudden change region is the fracture.
3. optical character recognition method according to claim 1, it is characterised in that the stroke polishing to the fracture
Also include：
When the region of fracture significantly increases region for the stroke width in the stroke of the fracture, by the region of fracture
The gray value in the stroke region outside the quasi- stroke region is set to the gray value with background area in the character block
Consistent value.
4. optical character recognition method according to claim 1, it is characterised in that enter described to the quasi- identification image
Before line character identification, in addition to：
Based on stroke width conversion, the noise in the character block is detected；
Eliminate the noise.
5. optical character recognition method according to claim 4, it is characterised in that described based on stroke width conversion, inspection
The noise surveyed in the character block includes：
The stroke width W of the character block is determined, character block i stroke width is equal to the first width average value and the second width is flat
/ 2nd of average sum, the first width average value is character block i average stroke width value, and character block i's is averaged
Stroke width value is equal to the average value of the stroke width value of all stroke area pixel points in character block i, and second width is put down
Average is the average value of the average stroke width value of all character blocks extracted from the original image, and i is natural number, pen
The stroke width value for drawing area pixel point becomes scaling method acquisition by stroke width；
Character stroke width error d is set；
Judge whether the stroke region that stroke width is in outside interval [W-d, W+d] in the character block meets following into bar of making an uproar
One of part：The stroke that is not consistent in the setting range in the stroke region with the stroke direction, along the stroke region
Stroke width change undulating value be more than setting fluctuation threshold or the stroke region stroke length be less than setting length threshold
Value；
When the stroke region that stroke width is in outside interval [W-d, W+d] in the character block meets described into one of condition of making an uproar
When, judge that the stroke region that stroke width is in outside interval [W-d, W+d] in the character block is noise.
6. optical character recognition method according to claim 4, it is characterised in that the elimination noise includes：
The gray value of the noise is set to the value consistent with the gray value of background area in the character block.
7. optical character recognition method according to claim 1 or 5, it is characterised in that the stroke width becomes scaling method
Including：
Rim detection is carried out to the character block, edge pixel point is found, the edge pixel point has a direction gradient value；
According to any edge pixel p gradient direction dpThe edge pixel point matched with edge pixel point p is searched along route r
Q, edge pixel point q gradient direction dqWith edge pixel point p gradient direction dpOn the contrary, wherein, r=p+n*dp, n be more than or
Equal to 0；
When finding the edge pixel point q, specify every on the edge pixel point p and edge pixel point q line
The stroke width value of individual pixel is the Euclidean distance between the edge pixel point p and edge pixel point q.
8. a kind of optical character recognition device, it is characterised in that including：
Extraction module, for extracting character block from original image；
Determining module, for being converted based on stroke width, determines the stroke being broken in the character block；
Polishing module, quasi- identification image is obtained for the stroke polishing to the fracture；
Identification module, for carrying out character recognition to the quasi- identification image, obtains the text message in the quasi- identification image；
Wherein, the polishing module includes：
Quasi- stroke area setting unit, if direction of the region of fracture along stroke where it and its place in the stroke of the fracture
The boundary pixel of stroke is respectively the first point of interface and the second point of interface, and the quasi- stroke area setting unit is used for will be described
It is equal to described first along stroke direction, the width of first point of interface between first point of interface and second point of interface to have a common boundary
The region of the average value of the stroke width value of point and the stroke width value of second point of interface sets the stroke region that is defined, its
In, the stroke width value of the stroke width value of first point of interface and second point of interface passes through stroke width converter unit
Obtain；
First gray scale adjusting unit, for the gray value in the quasi- stroke region to be set to and stroke region in the character block
The consistent value of gray value to obtain the quasi- identification image.
9. optical character recognition device according to claim 8, it is characterised in that the determining module includes：
Threshold setting unit, for setting breaking length threshold value；
Stroke direction determining unit, the direction of stroke where the stroke area pixel point for determining the character block；
It is mutated judging unit, the direction for judging the stroke where the stroke area pixel point, stroke region picture
It whether there is stroke width sudden change region, and the stroke width saltation zone between vegetarian refreshments and its next stroke area pixel point
The length in domain is less than the breaking length threshold value, and the stroke width sudden change region is white space, stroke width increment value is big
It is more than the region of the second change width threshold value of setting in the first change width threshold value of setting or stroke width decreasing value；
Stroke determining unit is broken, for when there is the stroke width sudden change region, determining the stroke area pixel point
Place stroke is the stroke of fracture, determines the region of fracture in the stroke that the stroke width sudden change region is the fracture.
10. optical character recognition device according to claim 8, it is characterised in that the polishing module also includes：
Second gray scale adjusting unit, for significantly increasing area when the stroke width in the stroke that the region of fracture is the fracture
During domain, the gray value that the stroke region outside the quasi- stroke region is in the region of fracture is set to and the character
The consistent value of the gray value of background area in block.
11. optical character recognition device according to claim 8, it is characterised in that before the identification module, also wrap
Include：
Noise detection module, for being converted based on stroke width, detects the noise in the character block；
Cancellation module, for eliminating the noise that the noise detection module is detected.
12. optical character recognition device according to claim 11, it is characterised in that the noise detection module includes：
Stroke width determining unit, the stroke width W for determining the character block, character block i stroke width is equal to first
/ 2nd of width average value and the second width average value sum, the first width average value is character block i average pen
Width value is drawn, character block i average stroke width value is equal to the stroke width value of all stroke area pixel points in character block i
Average value, the second width average value is the average stroke width value of all character blocks extracted from the original image
Average value, i is natural number, and the stroke width value of stroke area pixel point is obtained by stroke width converter unit；
Error setting unit, for setting character stroke width error d；
Into judging unit of making an uproar, for judging that stroke width is in the stroke region outside interval [W-d, W+d] in the character block
Whether meet following into one of condition of making an uproar：The pen not being consistent in the setting range in the stroke region with the stroke direction
Draw, along the stroke region stroke width change undulating value be more than setting fluctuation threshold or the stroke region stroke it is long
Degree is less than setting length threshold；
Identifying unit, for meeting institute when the stroke region that stroke width is in outside interval [W-d, W+d] in the character block
State into the condition of making an uproar for the moment, judge that stroke width is in the stroke region outside interval [W-d, W+d] to make an uproar in the character block
Sound.
13. optical character recognition device according to claim 11, it is characterised in that the cancellation module includes：
Gray scale is except unit of making an uproar, for the gray value of the noise to be set to the gray value one with background area in the character block
The value of cause.
14. the optical character recognition device according to claim 8 or 12, it is characterised in that the stroke width conversion is single
Member includes：
Rim detection subelement, for carrying out rim detection to the character block, finds edge pixel point, the edge pixel point
With a direction gradient value；
Coupling subelement, for the gradient direction d according to any edge pixel ppSearched and edge pixel point p phases along route r
The edge pixel point q matched somebody with somebody, edge pixel point q gradient direction dqWith edge pixel point p gradient direction dpOn the contrary, wherein, r=p
+n*dp, n is more than or equal to 0；
Subelement is specified, for when finding the edge pixel point q, specifying the edge pixel point p and the edge picture
The stroke width value of each pixel is the Europe between the edge pixel point p and edge pixel point q on vegetarian refreshments q line
Formula distance.
15. a kind of optical character recognition device, it is characterised in that including：
Processor；
Memory for storing processor-executable instruction；
Wherein, the processor is configured as performing and such as given an order：
Character block is extracted from original image；
Based on stroke width conversion, the stroke being broken in the character block is determined；
If direction of the region of fracture along stroke where it and the boundary pixel point of stroke where it in the stroke of the fracture
Not Wei the first point of interface and the second point of interface, will between first point of interface and second point of interface along described first have a common boundary
Stroke direction, the width of point are equal to the stroke width value of first point of interface and the stroke width value of second point of interface
The region of average value sets the stroke region that is defined, wherein, the stroke width value of first point of interface and second point of interface
Stroke width value pass through stroke width become scaling method obtain；
The gray value in the quasi- stroke region is set to the value consistent with the gray value in stroke region in the character block to obtain
Image is recognized to standard；Character recognition is carried out to the quasi- identification image, the text message in the quasi- identification image is obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410299362.5A CN104112130B (en) | 2014-06-26 | 2014-06-26 | optical character recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410299362.5A CN104112130B (en) | 2014-06-26 | 2014-06-26 | optical character recognition method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104112130A CN104112130A (en) | 2014-10-22 |
| CN104112130B true CN104112130B (en) | 2017-08-01 |
Family
ID=51708914
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410299362.5A Active CN104112130B (en) | 2014-06-26 | 2014-06-26 | optical character recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104112130B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104899588B (en) * | 2015-06-26 | 2018-12-11 | 小米科技有限责任公司 | Identify the method and device of the character in image |
| CN105447489B (en) * | 2015-11-13 | 2018-11-16 | 浙江传媒学院 | A kind of character of picture OCR identifying system and background adhesion noise cancellation method |
| CN106127118A (en) * | 2016-06-15 | 2016-11-16 | 珠海迈科智能科技股份有限公司 | A kind of English word recognition methods and device |
| CN110807416A (en) * | 2019-10-31 | 2020-02-18 | 国网湖北省电力有限公司电力科学研究院 | Digital instrument intelligent recognition device and method suitable for mobile detection device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4764972A (en) * | 1985-05-23 | 1988-08-16 | Nec Corporation | Continuous characters recognition system |
| US5029223A (en) * | 1990-02-02 | 1991-07-02 | International Business Machines Corporation | Constraint driven-on line recognition of handwritten characters and symbols |
| CN1438605A (en) * | 2003-03-14 | 2003-08-27 | 西安交通大学 | Beer-bottle raised character fetching-identifying hardware system and processing method |
| CN101320422A (en) * | 2008-06-06 | 2008-12-10 | 东莞市开心信息技术服务有限公司 | Normative decision method and apparatus for cross, connection and separation relationship of handwritten Chinese character strokes |
| CN101520851A (en) * | 2008-02-29 | 2009-09-02 | 富士通株式会社 | Character information identification device and method |
| CN102930277A (en) * | 2012-09-19 | 2013-02-13 | 上海珍岛信息技术有限公司 | Character picture verification code identifying method based on identification feedback |
-
2014
- 2014-06-26 CN CN201410299362.5A patent/CN104112130B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4764972A (en) * | 1985-05-23 | 1988-08-16 | Nec Corporation | Continuous characters recognition system |
| US5029223A (en) * | 1990-02-02 | 1991-07-02 | International Business Machines Corporation | Constraint driven-on line recognition of handwritten characters and symbols |
| CN1438605A (en) * | 2003-03-14 | 2003-08-27 | 西安交通大学 | Beer-bottle raised character fetching-identifying hardware system and processing method |
| CN101520851A (en) * | 2008-02-29 | 2009-09-02 | 富士通株式会社 | Character information identification device and method |
| CN101320422A (en) * | 2008-06-06 | 2008-12-10 | 东莞市开心信息技术服务有限公司 | Normative decision method and apparatus for cross, connection and separation relationship of handwritten Chinese character strokes |
| CN102930277A (en) * | 2012-09-19 | 2013-02-13 | 上海珍岛信息技术有限公司 | Character picture verification code identifying method based on identification feedback |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104112130A (en) | 2014-10-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106104575B (en) | Fingerprint template generation method and device | |
| CN105069786B (en) | Line detection method and device | |
| US20150332439A1 (en) | Methods and devices for hiding privacy information | |
| CN104731688B (en) | Point out the method and device of reading progress | |
| CN107239535A (en) | Similar pictures search method and device | |
| CN105528606A (en) | Region identification method and device | |
| CN106651955A (en) | Method and device for positioning object in picture | |
| EP3188078B1 (en) | Method and device for fingerprint identification | |
| CN105528602A (en) | Region identification method and device | |
| CN104735243B (en) | Contact list displaying method and device | |
| CN105528600A (en) | Region identification method and device | |
| CN105095881A (en) | Method, apparatus and terminal for face identification | |
| CN106228158A (en) | The method and apparatus of picture detection | |
| CN105159871A (en) | Text information detection method and apparatus | |
| KR20160048708A (en) | Recognition method and apparatus for communication message | |
| CN104112130B (en) | optical character recognition method and device | |
| CN106774803A (en) | Fingerprint identification method and device | |
| CN107480665A (en) | Character detecting method, device and computer-readable recording medium | |
| US20140267384A1 (en) | Display apparatus and control method thereof | |
| CN107742120A (en) | The recognition methods of bank card number and device | |
| CN105354560A (en) | Fingerprint identification method and device | |
| CN105447109A (en) | Key word searching method and apparatus | |
| CN105095860A (en) | Method and device for character segmentation | |
| CN107247936A (en) | Image-recognizing method and device | |
| CN105353938A (en) | Suspended short message display method and apparatus |
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 |