CN108830857A - A kind of adaptive Chinese character rubbings image binaryzation partitioning algorithm - Google Patents
A kind of adaptive Chinese character rubbings image binaryzation partitioning algorithm Download PDFInfo
- Publication number
- CN108830857A CN108830857A CN201810532680.XA CN201810532680A CN108830857A CN 108830857 A CN108830857 A CN 108830857A CN 201810532680 A CN201810532680 A CN 201810532680A CN 108830857 A CN108830857 A CN 108830857A
- Authority
- CN
- China
- Prior art keywords
- image
- algorithm
- rubbings
- chinese character
- segmentation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 49
- 230000003044 adaptive effect Effects 0.000 title claims abstract description 17
- 238000000638 solvent extraction Methods 0.000 title claims abstract description 14
- 238000000034 method Methods 0.000 claims abstract description 44
- 230000011218 segmentation Effects 0.000 claims abstract description 13
- 238000001914 filtration Methods 0.000 claims abstract description 10
- 238000005070 sampling Methods 0.000 claims abstract description 7
- 238000012545 processing Methods 0.000 claims description 10
- 238000005259 measurement Methods 0.000 claims description 6
- 230000000877 morphologic effect Effects 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 5
- 238000010191 image analysis Methods 0.000 claims description 4
- 238000013441 quality evaluation Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 238000006243 chemical reaction Methods 0.000 claims description 2
- 230000002596 correlated effect Effects 0.000 claims description 2
- 238000003708 edge detection Methods 0.000 claims description 2
- 238000005530 etching Methods 0.000 claims description 2
- 239000000284 extract Substances 0.000 claims description 2
- 238000007689 inspection Methods 0.000 claims description 2
- 230000007850 degeneration Effects 0.000 abstract description 14
- 238000005286 illumination Methods 0.000 abstract description 6
- 238000003709 image segmentation Methods 0.000 abstract description 4
- 238000001303 quality assessment method Methods 0.000 abstract description 3
- 238000012360 testing method Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000015556 catabolic process Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 210000003739 neck Anatomy 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 239000004575 stone Substances 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 238000003483 aging Methods 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 229910003460 diamond Inorganic materials 0.000 description 1
- 239000010432 diamond Substances 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 208000018769 loss of vision Diseases 0.000 description 1
- 231100000864 loss of vision Toxicity 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 230000004393 visual impairment Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20032—Median filtering
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Character Input (AREA)
Abstract
The invention discloses a kind of adaptive Chinese character rubbings image binaryzation partitioning algorithm, this method is divided into four steps：(1) it is pre-processed using median filtering；(2) red color components are extracted；(3) morphology operations are to find best background estimating；(4) Otsu divides bianry image.The invention belongs to Chinese character rubbings technical field of image segmentation, the stroke feature for retaining Chinese character enhances character details simultaneously, for the history rubbings image of degeneration, the image binaryzation adaptivenon-uniform sampling algorithm for the inhomogeneous illumination degeneration based on background estimating that the invention proposes a kind of, the novel place of the method proposed is to find the best background estimating based on blind/non-reference picture space quality assessment, the experimental results showed that this method can carry out more accurate Character segmentation to degeneration Chinese character.
Description
Technical field
The invention belongs to Chinese character rubbings technical field of image segmentation, in particular to a kind of adaptive Chinese character rubbings images two
Value partitioning algorithm.
Background technique
The important way that China retains oneself history culture is to write on epilithic memory --- inscription.Meanwhile rubbing is
The important component of chinese ancient book, it is the main source of people's study and research history.Rubbing digital library books are
The new way for carrying forward and passing on traditional Chinese art is the new approaches for protecting stone cultural artifact.On the other hand, existing ancient books rubbing is past
Toward Loss Of Vision quality.Over time, due to the influence of the factors such as its preservation condition, humidity, pollution, in rubbing
After duplication becomes digital picture, contrast is different, noise is big, due to background intensity enhancing etc., foreground and background be difficult by
It separates.Other than low contrast, ambient noise introduce texture stone toward contact, there are also the agings of paper material, therefore, from ancient rubbings
It is a challenging job that clean Chinese character is extracted in original image.But above-mentioned is also to carry out any further automatic text
The committed step etc. of shelves image analysis such as printed page analysis, character recognition etc..Luminance Distribution in rubbing image acquisition process is uneven
It is even, it will affect the quality of image.
In order to improve the quality of character cutting, inventor has studied a large amount of methods and techniques, and most important one is located in advance
Managing step is text binaryzation, file and picture is converted into bianry image from gray scale or color image, in background information by white
On the basis of pixel indicates, prospect is indicated by black picture element, the foreground and background of ancient times file and picture is separated.It is a kind of most simple and
Efficient image processing techniques can be used to separate the foreground and background layer of file and picture, be exactly thresholding, many can be classified as entirely
The threshold technology of office and local thresholding algorithm, multi-threshold method and auto-adaptive threshold technology.When image has in background and prospect
When having identical contrast, global threshold be it is preferred, their uniform illumination, target and background differs greatly.Part is adaptive
Answer threshold value for restoring the foreground pixel in file and picture.It in general, is one to one algorithm of degeneration rubbings image selection
Very difficult process.Due to there is complicated degeneration, it is many experimental results showed that, at traditional weak objective image uniform illumination
There is target and background separation be imperfect or the disadvantages such as target is lost, treatment effeciency is low for reason method.Therefore it inventors herein proposes
A kind of new adaptive algorithm handles rubbing image, i.e., utilizes natural image using blind/non-reference picture space quality assessment
Count (NSS) model framework locally normalization luminance factor, " non-natural " quantified using model parameter.It is low in order to correctly divide
Contrast color image utilizes rubbing structure using morphology Top-Hat operator and disc-shaped structure element and adaptive pixel
Between difference red component.In order to reduce noise, median filtering is applied to shadow correction image.
In general, the file and picture of degeneration is since the variation of ambient noise and contrast and brightness generates.Yin
The file and picture that shadow is degenerated is more common, because camera document is easier to be influenced by illumination variation.Existing many algorithms, it is intended to
Divide foreground and background in scan text document, but threshold value is the tool of the standard in certain form and another region, such as
The adaptive threshold of Bernsen is estimated according to the field of each pixel.Office is constructed using local maximum and minimum value
Portion's contrast image.Then, local threshold is determined using sliding window on this image.The prior art proposes one kind and is based on
Phase divides ancient times document image binaryzation model, develops a true Core Generator, referred to as PhaseGT, to simplify
With the generating process for accelerating real ancient times file and picture.Recently, it also proposed a kind of active profile evolution algorithmic in the prior art,
According to the inherent dimensional measurement of file and picture, picture contrast, i.e., by the local maximum of image and minimum value, for automatically raw
At our movable contour model, initialisation image；Finally, average threshold also can produce and final binaryzation.As invention
Observed by people, most of binarization methods be based on the intuitive observation to tonal gradation between character and background, but regardless of
The adaptively selected threshold value of the file and picture of degeneration.In order to overcome these difficulties, this programme proposes a kind of adaptive method,
Carry out separating character from the file and picture of degeneration using different methods.
Summary of the invention
To solve above-mentioned existing problem, retains the stroke feature of Chinese character while enhancing character details, for the history of degeneration
The image binaryzation adaptivenon-uniform sampling of rubbings image, the inhomogeneous illumination degeneration based on background estimating that the invention proposes a kind of is calculated
Method, the novel place of the method proposed are to find the best background based on blind/non-reference picture space quality assessment
Estimation, this method are divided into four steps：(1) it is pre-processed using median filtering；(2) red color components are extracted；(3) morphology operations
To find best background estimating；(4) Otsu divides bianry image.The experimental results showed that this method can carry out degeneration Chinese character
More accurate Character segmentation.
The technical solution adopted by the present invention is as follows：A kind of adaptive Chinese character rubbings image binaryzation partitioning algorithm uses
Color image develops a kind of robust algorithm, for dividing Chinese rubbing image from background, includes the following steps：
1) it is handled again using median filtering, median filtering allows a large amount of high spatial frequency details to pass through, while having very much
Eliminate the noise being less than on the image of half pixel in smooth neighborhood in effect ground；
2) red color components are extracted；
3) morphological images processing operation, so as to if it find that minimum BRISQUE, then can find out the optimum diameter of disk
Thr*；
4) the segmentation bianry image of Otsu is used.
Further, morphology described in step 3) is mathematical morphology, carries out image procossing using mathematical morphology
Basic thought be using have effigurate structural element (basic element with certain planform, as rectangle, circle or
Diamond shape etc.) target image is detected, it is analyzed by the validity to structural element in image target area and fill method,
The relevant information of image aspects and structure is obtained, and realizes the purpose of image analysis and identification using them.
Further, the structural element is a key point of morphological images processing, and different structural elements determines
The analysis and processing of various geological informations, also determine the calculation amount in data conversion process, therefore to structural elements in image
The analysis of element is the important content of Image Edge-Detection；The size and planform of structural element can all influence image border inspection
It surveys；Small scale structures element has weaker noise removal capability, but they can detecte accurate edge details；Coarse scale structures member
Element has stronger noise removal capability, but the edge detected is more coarse；Importantly, structural element of different shapes is not to
There is different processing capacities at edge with image；Wherein, gray level image can be regarded as one group of two-dimensional points, expansion and etching operation
It can be expressed as follows：
Top-hat algorithm can be divided into Top-hat algorithm and Bot- according to the different components of open operation and closed procedure
Top-hat algorithm is applied to image and is expressed as TH by hat algorithm：
Top-hat algorithm is applied to image and is expressed as BH：
BH (x, y)=(f ⊙ S-f) (x, y) (4)
In equation, f (x, y) is original-gray image, and S (x, y) is structural element, and Top-hat transformation passes through original graph
As extracting foreground information with the difference between its opening operation, and Bot-hat transformation passes through original image and its closed procedure
Between difference inhibit background information.
Further, BRISQUE described in step 3) is a kind of general no with reference to figure based on spatial image statistical nature
Image quality amount assessment algorithm, the algorithm are based on following theoretical premise：Natural image has certain regularity, the visual signature of human eye
As rule develops, in the prior art, Anish Mittal and other it was discovered by researchers that natural image in spatial domain
Normalization luminance factor has statistical property and meets unit Gaussian Profile.This function is influenced by image fault, different mistakes
Very to being assigned different influences.Based on the above research achievement, we have proposed a kind of based on spatial domain statistical nature
BRISQUE non-reference picture quality evaluation algorithm.The gray level image for being M*N for given size, the brightness of each pixel are returned
One change coefficient meets as follows：
In formula：I=1,2 ..., M；J=1,2 ..., N；C is constant, c=1；K=L=3；I M=... 1,2,；j N
=... 1,2,；C is a constant, c=1；K=L=3；μ (i, j) and σ (i, j) is average and standard deviation；ω=
{ω}kj∣ k=-K ,-K+1 ..., K, L=-L ,-L+1 ..., L is the sampling and standardization of dimensional Gaussian equation；BRISQUE is calculated
Method uses brightness normalization coefficient as quality correlated characteristic to assess picture quality.Compared with other non-reference quality evaluations,
The use of characteristics of image eliminates the demand to various complex transformations.Therefore, the algorithm similar in the precision under the premise of there is meter
Calculate simple, time saving advantage.On the other hand, the decorrelative transformation of brightness of image has ignored brightness to test image quality
It influences.
Further, it includes binarization segmentation algorithm, binarization segmentation in the segmentation bianry image of Otsu that step 4), which uses,
Layering measurement, false positive rate are carried out including tri- Jaccard coefficient, false positive rate FPR and false negative rate FNR parameters in algorithm
FPR shows less divided degree, and false negative rate FNR shows over-segmentation degree；Jaccard coefficient measures between finite sample collection
Similitude, and the size for being defined as intersection calculates bianry image for bianry image divided by the size of the union of sample set
The friendship of A and B divided by A and B's and, Jaccard coefficient can be used following formula calculating：
False positive rate FPR and false negative rate FNR are defined as follows：
Wherein FP is the quantity of wrong report, the white in true picture and the black in binary image, and FN is false negative
Quantity, TN are the quantity of true negative, and N=FP+TN is negative total quantity.
The present invention is obtained and is had the beneficial effect that using the above scheme：The stroke feature that the present invention retains Chinese character enhances word simultaneously
Details is accorded with, for the history rubbings image of degeneration, proposes a kind of image two that the inhomogeneous illumination based on background estimating is degenerated
Value adaptivenon-uniform sampling algorithm, the novel place of the method proposed are to find one based on blind/non-reference picture space matter
The best background estimating of assessment is measured, this method is divided into four steps：(1) it is pre-processed using median filtering；(2) extract it is red at
Point；(3) morphology operations are to find best background estimating；(4) Otsu divides bianry image.The experimental results showed that this method energy
It is enough that more accurate Character segmentation is carried out to degeneration Chinese character.
Detailed description of the invention
Fig. 1 is the method for the present invention flow chart；
Fig. 2 is the rubbing image of conventional Chinese rubbings image；
Fig. 3 is the histogram of Fig. 2；
Fig. 4 is the target function value figure of Fig. 2；
Fig. 5 is image segmentation result comparative diagram.
Specific embodiment
The technical scheme in the embodiments of the invention will be clearly and completely described below, it is clear that described implementation
Example is only a part of the embodiments of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, this field is common
Technical staff's all other embodiment obtained without making creative work belongs to the model that the present invention protects
It encloses.
A kind of embodiment, adaptive Chinese character rubbings image binaryzation partitioning algorithm of the present embodiment, is opened using color image
A kind of robust algorithm has been sent out, for dividing Chinese rubbing image from background, has been included the following steps：
1) it is handled again using median filtering, median filtering allows a large amount of high spatial frequency details to pass through, while having very much
Eliminate the noise being less than on the image of half pixel in smooth neighborhood in effect ground；
2) red color components are extracted；
3) morphological images processing operation, so as to if it find that minimum BRISQUE, then can find out the optimum diameter of disk
Thr*；
4) the segmentation bianry image of Otsu is used.
The flow chart of the dividing method proposed is as shown in Fig. 1.
Experimental analysis and as a result,
In order to assess and test proposed method, this programme, which uses, comes from University of California-Berkeley East Asia
The Chinese rubbings image in library database (http://ucblibrary4.berkeley.edu:8088/xtf3/ search？Rmode=stonerubbings&identifier=&title=&name=&text=&date=&
Startdate=15&subje ct=&height=&width=&material=&script=&enc_provenance
=).This programme Jaccard similarity factor, FPR (false positive rate) and FNR (false negative rate) and classical Otsu algorithm phase
Than to assess performance of the system of this programme in performance and quality.As in Figure 2-4, the rubbing of conventional Chinese rubbings image
Image, rubbing image is partially dark, and histogram shows that picture contrast is very low, history archive image quality decrease.Estimation prospect it
Before, it is necessary to noise is removed using median filter.Median filter successively considers each pixel in image and checks its neck
Domain, to determine whether it represents its field.The neighborhood is chosen as the square of 3 × 3 pixels.Compared with image size, this is one
A very small neighborhood, final image size are 371 × 1260 pixels.Later, function imopen is used in Matlab,
Morphology opening operation is executed in estimation prospect.
The binarization method of above-mentioned introduction is applied to our test set, the test set by with several degradations and
The old Chinese rubbings file and picture composition of structural complexity.We introduce in the present embodiment receives former method applied to us
The result of the image of hiding.For objectively evaluating, we use Jaccard coefficient, false positive rate (FPR) and false negative rate (FNR)
Three parameters carry out layering measurement, and FPR shows less divided degree, and FNR shows over-segmentation degree.Jaccard coefficient measurement has
The similitude between sample set is limited, and is defined as the size of intersection divided by the size of the union of sample set, for bianry image,
It calculate bianry image A and B friendship divided by A and B's and.Following formula calculating can be used in Jaccard coefficient：
False positive rate FPR and false negative rate FNR are defined as follows：
Wherein FP is the quantity of wrong report, the white in true picture and the black in binary image, and FN is false negative
Quantity, TN are the quantity of true negative, and N=FP+TN is negative total quantity.
The performance of partitioning algorithm is as shown in table 1.
The quantitative measurement results that table 1 is divided
(d) is the Chinese character segmentation using the acquisition of this programme adaptivenon-uniform sampling algorithm as a result, (c) is by obtaining in Fig. 5 in Fig. 5
Ostu' partitioning algorithm obtain as a result, (d) is substantially true bianry image in Fig. 5, refer to eliminating all noises manually
With the practical bianry image of the deterioration factor.In table 1, it can be noted that the optimum of low contrast file and picture is to pass through this
What scheme method obtained.In table 1, Jaccard coefficient is significant to be higher than Ostu method.The global threshold method mistake of OTSU has divided one
A little text pixels, while dark background pixel is mistakenly classified as text pixel.The experimental results showed that this programme was proposed
Background elimination algorithm ratio Ostu method may be implemented more accurately to repair for various Chinese rubbings images.For low contrast
Chinese rubbings image, it shows good.
But since threshold value is global application, for the threshold value of certain weak person's handwritings, person's handwriting is caused to be possible to be destroyed.
To sum up, the features such as Chinese rubbings image that rubbing obtains has fuzzy details more, and effect is poor, therefore at traditional place
More details may be lost during reason.Pretreatment is an important stage in image procossing, particularly in China ancient books
In the case that image segmentation is applied.A kind of efficient Image Pretreatment Algorithm will improve the accuracy of partitioning algorithm and reduce mistake
Classification.The invention proposes a kind of adaptivenon-uniform sampling algorithms of rubbings image binaryzation for degeneration.It is subjective and objectively comment
Valence method is used to judge the efficiency of our algorithms.The experimental results showed that estimating that image background is adaptive by morphological operation
It should select to find the optimum diameter of disk.But since our background estimating algorithm does not account for the photograph in different scenes
Degree relationship, so in the case where scene significant change, it is possible to introduce the slight flashing of other application.In the future, this method will
It is tested in OCR (optical character identification) application program, to test readability of the proposed method in degradation document.
In addition, the present invention obtains National Natural Science Foundation of China (NSFC) (61472173), Jiangxi Province's Natural Science Fund In The Light
(20161BAB202042), the support energetically of Education Commission of Jiangxi Province Funded Projects (GJJ151134).
The above description is only an embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright description is applied directly or indirectly in other relevant technology necks
Domain is included within the scope of the present invention.
Claims (5)
1. a kind of adaptive Chinese character rubbings image binaryzation partitioning algorithm, which is characterized in that develop one using color image
Kind robust algorithm includes the following steps for dividing Chinese rubbing image from background：
1) it is handled again using median filtering, median filtering allows a large amount of high spatial frequency details to pass through, while effectively
It eliminates in smooth neighborhood less than the noise on the image of half pixel；
2) red color components are extracted；
3) morphological images processing operation, so as to if it find that minimum BRISQUE, then can find out the optimum diameter Thr* of disk；
4) the segmentation bianry image of Otsu is used.
2. the adaptive Chinese character rubbings image binaryzation partitioning algorithm of one kind according to claim 1, which is characterized in that step
It is rapid 3) described in morphology be mathematical morphology, using mathematical morphology carry out image procossing basic thought be using have one
The structural elements of setting shape usually detect target image, pass through the validity to structural element in image target area and fill method
It is analyzed, obtains the relevant information of image aspects and structure, and realize the purpose of image analysis and identification using them.
3. the adaptive Chinese character rubbings image binaryzation partitioning algorithm of one kind according to claim 2, which is characterized in that institute
The key point that structural element is morphological images processing is stated, different structural elements determines various geological informations in image
Analysis and processing, also determine the calculation amount in data conversion process, therefore be image border inspection to the analysis of structural element
The important content of survey；The size and planform of structural element can all influence Image Edge-Detection；Small scale structures element has
Weaker noise removal capability, but they can detecte accurate edge details；Coarse scale structures element has stronger noise removal capability,
But the edge detected is more coarse；Importantly, structural element of different shapes have to the edge of different images it is different
Processing capacity；Wherein, gray level image can be regarded as one group of two-dimensional points, and expansion and etching operation can be expressed as follows：
(f ⊙ S) (x, y)=min g (x-k, y-l) | (k, l) ∈ S } (2)
Top-hat algorithm can be divided into Top-hat algorithm according to the different components of open operation and closed procedure and Bot-hat is calculated
Top-hat algorithm is applied to image and is expressed as TH by method：
Top-hat algorithm is applied to image and is expressed as BH：
BH (x, y)=(f ⊙ S-f) (x, y) (4)
In equation, f (x, y) is original-gray image, and S (x, y) is structural element, Top-hat transformation by original image with
Difference between its opening operation extracts foreground information, and Bot-hat transformation is by between original image and its closed procedure
Difference inhibit background information.
4. the adaptive Chinese character rubbings image binaryzation partitioning algorithm of one kind according to claim 1, which is characterized in that step
It is rapid 3) described in BRISQUE be a kind of general non-reference picture quality evaluation algorithm based on spatial image statistical nature, for
Given size is the gray level image of M*N, and the brightness normalization coefficient of each pixel meets as follows：
In formula：I=1,2 ..., M；J=1,2 ..., N；C is constant, c=1；K=L=3；I M=... 1,2 ...,；J N=...
1,2,…,；C is a constant, c=1；K=L=3；μ (i, j) and σ (i, j) is average and standard deviation；ω={ ω }kj
∣ k=-K ,-K+1 ..., K, L=-L ,-L+1 ..., L is the sampling and standardization of dimensional Gaussian equation；BRISQUE algorithm uses
Brightness normalization coefficient assesses picture quality as quality correlated characteristic.
5. the adaptive Chinese character rubbings image binaryzation partitioning algorithm of one kind according to claim 1, which is characterized in that step
Include binarization segmentation algorithm in the rapid segmentation bianry image for 4) using Otsu, includes Jaccard system in binarization segmentation algorithm
Tri- number, false positive rate FPR and false negative rate FNR parameters carry out layering measurement, and false positive rate FPR shows less divided degree, false
Negative rate FNR shows over-segmentation degree；Jaccard coefficient measures the similitude between finite sample collection, and is defined as phase
The size of friendship divided by the union of sample set size, for bianry image, calculate the friendship of bianry image A and B divided by A and B's and,
Following formula calculating can be used in Jaccard coefficient：
False positive rate FPR and false negative rate FNR are defined as follows：
Wherein FP is the quantity of wrong report, the white in true picture and the black in binary image, and FN is the quantity of false negative,
TN is the quantity of true negative, and N=FP+TN is negative total quantity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810532680.XA CN108830857B (en) | 2018-05-29 | 2018-05-29 | Self-adaptive Chinese character copy label image binarization segmentation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810532680.XA CN108830857B (en) | 2018-05-29 | 2018-05-29 | Self-adaptive Chinese character copy label image binarization segmentation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108830857A true CN108830857A (en) | 2018-11-16 |
| CN108830857B CN108830857B (en) | 2021-04-23 |
Family
ID=64146270
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810532680.XA Expired - Fee Related CN108830857B (en) | 2018-05-29 | 2018-05-29 | Self-adaptive Chinese character copy label image binarization segmentation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108830857B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109668912A (en) * | 2018-12-05 | 2019-04-23 | 余姚德诚科技咨询有限公司 | With bed X-ray machine quality inspection mechanism |
| CN110163212A (en) * | 2019-04-08 | 2019-08-23 | 天津大学 | A kind of text cutting method in rubbings image |
| CN111462144A (en) * | 2020-03-30 | 2020-07-28 | 南昌工程学院 | Image segmentation method for rapidly inhibiting image fuzzy boundary based on rough set |
| CN112287933A (en) * | 2019-12-20 | 2021-01-29 | 中北大学 | Method and system for removing character interference of X-ray image of automobile hub |
| CN113628235A (en) * | 2021-05-10 | 2021-11-09 | 哈尔滨工业大学 | Self-adaptive threshold segmentation signature image binary processing method based on PDE |
| CN113989814A (en) * | 2021-11-23 | 2022-01-28 | 腾讯科技（深圳）有限公司 | Image generation method and device, computer equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004227355A (en) * | 2003-01-24 | 2004-08-12 | Oki Electric Ind Co Ltd | Scanner system, sheet for setting threshold, and binary image forming method |
| CN105160678A (en) * | 2015-09-02 | 2015-12-16 | 山东大学 | Convolutional-neural-network-based reference-free three-dimensional image quality evaluation method |
| CN107085726A (en) * | 2017-01-04 | 2017-08-22 | 安阳师范学院 | Oracle bone rubbing individual character localization method based on multi-method denoising and connected component analysis |
| CN107909071A (en) * | 2017-12-19 | 2018-04-13 | 安徽理工大学 | A kind of license plate locating method based on mathematical morphology |
-
2018
- 2018-05-29 CN CN201810532680.XA patent/CN108830857B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004227355A (en) * | 2003-01-24 | 2004-08-12 | Oki Electric Ind Co Ltd | Scanner system, sheet for setting threshold, and binary image forming method |
| CN105160678A (en) * | 2015-09-02 | 2015-12-16 | 山东大学 | Convolutional-neural-network-based reference-free three-dimensional image quality evaluation method |
| CN107085726A (en) * | 2017-01-04 | 2017-08-22 | 安阳师范学院 | Oracle bone rubbing individual character localization method based on multi-method denoising and connected component analysis |
| CN107909071A (en) * | 2017-12-19 | 2018-04-13 | 安徽理工大学 | A kind of license plate locating method based on mathematical morphology |
Non-Patent Citations (1)
| Title |
|---|
| ZHI-KAI HUANG ET AL.: "Binarization Chinese Rubbing Images Using Gaussian Mixture Model", 《ICIC 2015》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109668912A (en) * | 2018-12-05 | 2019-04-23 | 余姚德诚科技咨询有限公司 | With bed X-ray machine quality inspection mechanism |
| CN109668912B (en) * | 2018-12-05 | 2022-01-04 | 马亮 | Quality inspection mechanism for X-ray machine |
| CN110163212A (en) * | 2019-04-08 | 2019-08-23 | 天津大学 | A kind of text cutting method in rubbings image |
| CN110163212B (en) * | 2019-04-08 | 2023-05-23 | 天津大学 | Text cutting method in inscription image |
| CN112287933A (en) * | 2019-12-20 | 2021-01-29 | 中北大学 | Method and system for removing character interference of X-ray image of automobile hub |
| CN111462144A (en) * | 2020-03-30 | 2020-07-28 | 南昌工程学院 | Image segmentation method for rapidly inhibiting image fuzzy boundary based on rough set |
| CN111462144B (en) * | 2020-03-30 | 2023-07-21 | 南昌工程学院 | Image segmentation method for rapidly inhibiting image fuzzy boundary based on rough set |
| CN113628235A (en) * | 2021-05-10 | 2021-11-09 | 哈尔滨工业大学 | Self-adaptive threshold segmentation signature image binary processing method based on PDE |
| CN113989814A (en) * | 2021-11-23 | 2022-01-28 | 腾讯科技（深圳）有限公司 | Image generation method and device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108830857B (en) | 2021-04-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108830857A (en) | A kind of adaptive Chinese character rubbings image binaryzation partitioning algorithm | |
| Senthilkumaran et al. | Image segmentation by using thresholding techniques for medical images | |
| CN110490914B (en) | Image fusion method based on brightness self-adaption and significance detection | |
| Gatos et al. | ICDAR 2009 document image binarization contest (DIBCO 2009) | |
| Pratikakis et al. | ICFHR 2012 competition on handwritten document image binarization (H-DIBCO 2012) | |
| CN111104943B (en) | Color image region-of-interest extraction method based on decision-level fusion | |
| US10453195B2 (en) | Method of detecting tissue area of interest in digital pathology imaging by executing computer-executable instructions stored on a non-transitory computer-readable medium | |
| CN106934386B (en) | A kind of natural scene character detecting method and system based on from heuristic strategies | |
| Alkoffash et al. | A survey of digital image processing techniques in character recognition | |
| CN113962976B (en) | Quality evaluation method for pathological slide digital image | |
| Dixit et al. | Image texture analysis-survey | |
| CN109949294A (en) | A kind of fracture apperance figure crack defect extracting method based on OpenCV | |
| CN115082776A (en) | Electric energy meter automatic detection system and method based on image recognition | |
| CN109726660A (en) | A kind of remote sensing images ship identification method | |
| CN116596899A (en) | Method, device, terminal and medium for identifying circulating tumor cells based on fluorescence image | |
| Mohan et al. | An intelligent recognition system for identification of wood species | |
| CN110276260B (en) | Commodity detection method based on depth camera | |
| Widynski et al. | A contrario edge detection with edgelets | |
| Yahya et al. | Image enhancement background for high damage Malay manuscripts using adaptive threshold binarization | |
| Zambanini et al. | Robust automatic segmentation of ancient coins | |
| Vu et al. | Automatic extraction of text regions from document images by multilevel thresholding and k-means clustering | |
| Messaoud et al. | Region based local binarization approach for handwritten ancient documents | |
| Saha et al. | Npix2Cpix: A GAN-based Image-to-Image Translation Network with Retrieval-Classification Integration for Watermark Retrieval from Historical Document Images | |
| Jyoti et al. | Document image binarization technique for degraded document images by using morphological operators | |
| CN109214363A (en) | A kind of substation's worker's face identification method based on YCbCr and connected component analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210423 |