CN108830270A - Correctly divide the localization method of the language of the Manchus word central axes of each identification to language of the Manchus word - Google Patents
Correctly divide the localization method of the language of the Manchus word central axes of each identification to language of the Manchus word Download PDFInfo
- Publication number
- CN108830270A CN108830270A CN201810580666.7A CN201810580666A CN108830270A CN 108830270 A CN108830270 A CN 108830270A CN 201810580666 A CN201810580666 A CN 201810580666A CN 108830270 A CN108830270 A CN 108830270A
- Authority
- CN
- China
- Prior art keywords
- language
- wdmax
- wdmin
- manchus
- image
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Character Discrimination (AREA)
Abstract
This divisional application discloses the localization method that a kind of pair of language of the Manchus word correctly divides the language of the Manchus word central axes of each identification, belongs to word processing field, for solves the problems, such as improve recognition accuracy, main points be include S1. word image pre-process；S2. the width of language of the Manchus word central axes is determined；S3. language of the Manchus word central axes right boundary position is determined, effect is to improve recognition accuracy.
Description
The application is application number 2015106293443, applying date 2015-09-29, denomination of invention " language of the Manchus word central axes
Localization method " divisional application.
Technical field
The present invention relates to a kind of literal processing methods, the especially a kind of localization method of language of the Manchus word central axes.
Background technique
With the development of science and technology, computer has been able to identify the text of a variety of picture formats and recognized
(such as English, Japanese), the one kind of the language of the Manchus as universal language can not also accomplish for Computer Automatic Recognition.If meter can be allowed
Calculation machine accomplishes automatic identification, the identification language of the Manchus, then a large amount of Manchu literatures can be carried out electronic disposal, this is for Manchu literatures
Research and preservation or even the research of Manchu culture and preservation suffer from highly important meaning, and to realize this target,
It is positioned firstly the need of the central axes to the language of the Manchus.
Summary of the invention
The present invention is to propose that a kind of accuracy rate is high in order to solve above-mentioned deficiency present in the prior art, help to calculate
Machine is correctly divided the localization method of the language of the Manchus word central axes of each identification to language of the Manchus word.
The technical solution of the invention is as follows：A kind of localization method of language of the Manchus word central axes, it is characterised in that：Described
Method successively by the pretreatment of language of the Manchus word image, the width for determining language of the Manchus word central axes, determine language of the Manchus word central axes or so
Three step compositions of boundary position, wherein：
The pretreated step of language of the Manchus word image is：Language of the Manchus document is scanned, obtains the color image of language of the Manchus document, and to full
The color image of document carries out binaryzation, and the image of each column language of the Manchus document is extracted using vertical projection method, obtains column image,
And language of the Manchus word image is extracted using horizontal projection method to column image, then language of the Manchus word image is carried out at place normalization
Reason, the step of normalized are the coordinate positions for finding four points in language of the Manchus word image outermost upper and lower, left and right, with this four
A point is boundary, redefines the size of language of the Manchus word image, and the two-value language of the Manchus after respectively indicating normalized with I, W, H
The height number of word image, the width numerical value of two-value language of the Manchus word image and two-value language of the Manchus word image,
The step of determining language of the Manchus word axis line width be：Two-value language of the Manchus word image I is carried out using mathematical morphology
Processing, specific method be use length for 2 horizontal direction linear type morphological structuring elements, to two-value language of the Manchus word image I
Corrosion treatment is carried out, obtains the corrosion image Ie of I, and seek the longitudinal edge figure of two-value language of the Manchus word image I Yu corrosion image Ie
Picture Iv, i.e. Iv=I-Ie extract the longitudinal edge of language of the Manchus word, then carry out upright projection, system to longitudinal edge image Iv
The number of black picture element included in each column of longitudinal edge image Iv is counted, upright projection sequence is indicated with vp1, amount of orientation
Unduplicated element in vp1, and arranged according to the descending mode of numerical value from big to small, new vector v p2 is obtained, and take
Preceding 3 elements in vp2 are respectively labeled as ind1, ind2 and ind3, indicate that the width of central axes, wdmin indicate minimum with wd
Axis line width, assignment wdmin=0, wdmax indicate that maximum axis line width, T are the threshold value of maximum axis line width, assignment T
=1/3, then maximum axis line width wdmax is the value after being rounded downwards to the result of T × W, i.e.,
Defining wd1, wd2, wd3 respectively is 3 candidate axis line widths, and wd indicates the axis line width finally determined,
Middle ∣ -1 wd1=∣ ind1-ind2 ∣ -1, wd2=∣ ind1-ind3 ∣ -1, wd3=∣ ind2-ind3,
If (wdmin<wd1<) and (wdmin wdmax<wd2<) and (wdmin wdmax<wd3<Wdmax), then wd
=wd1,
If (wdmin<wd1<) and (wdmin wdmax<wd2<) and (wd3 wdmax<=wdmin or wd3>=
Wdmax), then the biggish output valve as wd of wd1 and wd2 intermediate value, i.e. wd=max (wd1, wd2) are taken,
If (wdmin<wd1<) and (wdmin wdmax<wd3<) and (wd2 wdmax<=wdmin or wd2>=
Wdmax), then the biggish output valve as wd of wd1 and wd3 intermediate value, i.e. wd=max (wd1, wd3) are taken,
If (wdmin<wd2<) and (wdmin wdmax<wd3<) and (wd1 wdmax<=wdmin or wd1>=
Wdmax), then the biggish output valve as wd of wd2 and wd3 intermediate value, i.e. wd=max (wd2, wd3) are taken,
If (wdmin<wd1<) and (wd2 wdmax<=wdmin or wd2>=wdmax) and (wd3<=wdmin
Or wd3>=wdmax), then wd=wd1,
If (wdmin<wd2<) and (wd1 wdmax<=wdmin or wd1>=wdmax) and (wd3<=wdmin
Or wd3>=wdmax), then wd=wd2,
If (wdmin<wd3<) and (wd1 wdmax<=wdmin or wd1>=wdmax) and (wd2<=wdmin
Or wd2>=wdmax), then wd=wd3,
If (wd1<=wdmin or wd1>=wdmax) and (wd2<=wdmin or wd2>=wdmax) and
(wd3<=wdmin or wd3>=wdmax), then wd=wdmax,
The step of determining language of the Manchus word central axes right boundary position be：Indicate that central axes are most left with baseline_left
The boundary of side scans by column longitudinal edge image Iv according to sequence from left to right, and scanning range is from the 1st column to W-
Wd+1 column, if being currently the i-th column, then 1≤i≤(W-wd+1), indicates the i-th column to the i-th+wd-1 black pixel point arranged with TB
Number, then baseline_left be with column corresponding to maximum TB value,
The boundary of the central axes rightmost side is indicated with baseline_right, then baseline_right=baseline_
Left+wd-1,
Therefore in two-value language of the Manchus word image I, baseline_left is the left margin of the language of the Manchus word central axes,
Baseline_right is the right margin of the language of the Manchus word central axes, and wd is the width of the language of the Manchus word central axes, to position
The central axes of the language of the Manchus word out.
The present invention compared with the existing technology, has the following advantages that：
Using the method disclosed in the present, central axes (including the axis of language of the Manchus word can be quickly and accurately determined
The right boundary position of line and the width of central axes), and the determination of the central axes of language of the Manchus word, it is that computer can correctly be divided
And identifying the premise and basis of language of the Manchus text, the appearance of this kind of method allows the computer of the language of the Manchus to identify and electronic disposal obtains
Further development, the research and preservation of research and preservation and Manchu culture for Manchu literatures all have great significance
, it can be said that this method has the advantages that a variety of, particularly suitable for promoting and applying in the art, market prospects are extensive,
Significance of scientific research is great.
Specific embodiment
A specific embodiment of the invention is described below.
A kind of localization method of language of the Manchus word central axes according to the pretreatment of language of the Manchus word image, determines language of the Manchus word axis
The width of line determines that language of the Manchus word central axes right boundary position three step is operated,
The pretreated step of language of the Manchus word image is：Language of the Manchus document is scanned, obtains the color image of language of the Manchus document, and to full
The color image of document carries out binaryzation, and the image of each column language of the Manchus document is extracted using vertical projection method, obtains column image,
And language of the Manchus word image is extracted using horizontal projection method to column image, then language of the Manchus word image is carried out at place normalization
Reason, the step of normalized are the coordinate positions for finding four points in language of the Manchus word image outermost upper and lower, left and right, with this four
A point is boundary, redefines the size of language of the Manchus word image, and the two-value language of the Manchus after respectively indicating normalized with I, W, H
The height number of word image, the width numerical value of two-value language of the Manchus word image and two-value language of the Manchus word image,
The step of determining language of the Manchus word axis line width be：Two-value language of the Manchus word image I is carried out using mathematical morphology
Processing, specific method be use length for 2 horizontal direction linear type morphological structuring elements, to two-value language of the Manchus word image I
Corrosion treatment is carried out, obtains the corrosion image Ie of I, and seek the longitudinal edge figure of two-value language of the Manchus word image I Yu corrosion image Ie
Picture Iv, i.e. Iv=I-Ie extract the longitudinal edge of language of the Manchus word, then carry out upright projection, system to longitudinal edge image Iv
The number of black picture element included in each column of longitudinal edge image Iv is counted, upright projection sequence is indicated with vp1, amount of orientation
Unduplicated element in vp1, and arranged according to the descending mode of numerical value from big to small, new vector v p2 is obtained, and take
Preceding 3 elements in vp2 are respectively labeled as ind1, ind2 and ind3, indicate that the width of central axes, wdmin indicate minimum with wd
Axis line width, assignment wdmin=0, wdmax indicate that maximum axis line width, T are the threshold value of maximum axis line width, assignment T
=1/3, then maximum axis line width wdmax is the value after being rounded downwards to the result of T × W, i.e.,
Defining wd1, wd2, wd3 respectively is 3 candidate axis line widths, and wd indicates the axis line width finally determined,
Middle ∣ -1 wd1=∣ ind1-ind2 ∣ -1, wd2=∣ ind1-ind3 ∣ -1, wd3=∣ ind2-ind3,
If (wdmin<wd1<) and (wdmin wdmax<wd2<) and (wdmin wdmax<wd3<Wdmax), explanation
Wd1, wd2, wd3 are within the scope of rational width, then wd=wd1,
If (wdmin<wd1<) and (wdmin wdmax<wd2<) and (wd3 wdmax<=wdmin or wd3>=
Wdmax), illustrate wd1, wd2 within the scope of rational width, and wd3 is within the scope of rational width, then take wd1 and wd2 intermediate value compared with
A big output valve as wd, i.e. wd=max (wd1, wd2),
If (wdmin<wd1<) and (wdmin wdmax<wd3<) and (wd2 wdmax<=wdmin or wd2>=
Wdmax), illustrate wd1, wd3 within the scope of rational width, and wd2 is within the scope of rational width, then take wd1 and wd3 intermediate value compared with
A big output valve as wd, i.e. wd=max (wd1, wd3),
If (wdmin<wd2<) and (wdmin wdmax<wd3<) and (wd1 wdmax<=wdmin or wd1>=
Wdmax), illustrate wd2, wd3 within the scope of rational width, and wd1 is within the scope of rational width, then take wd2 and wd3 intermediate value compared with
A big output valve as wd, i.e. wd=max (wd2, wd3),
If (wdmin<wd1<) and (wd2 wdmax<=wdmin or wd2>=wdmax) and (wd3<=wdmin
Or wd3>=wdmax), illustrate wd1 within the scope of rational width, and wd2, wd3 be within the scope of rational width, then wd=
Wd1,
If (wdmin<wd2<) and (wd1 wdmax<=wdmin or wd1>=wdmax) and (wd3<=wdmin
Or wd3>=wdmax), illustrate wd2 within the scope of rational width, and wd1, wd3 be within the scope of rational width, then wd=
Wd2,
If (wdmin<wd3<) and (wd1 wdmax<=wdmin or wd1>=wdmax) and (wd2<=wdmin
Or wd2>=wdmax), illustrate wd3 within the scope of rational width, and wd1, wd2 not within the scope of rational width then wd=
Wd3,
If (wd1<=wdmin or wd1>=wdmax) and (wd2<=wdmin or wd2>=wdmax) and
(wd3<=wdmin or wd3>=wdmax), illustrate wd1, wd2, wd3 not within the scope of rational width, then wd=wdmax,
It is specific to walk after the width that language of the Manchus word central axes have been determined, then come the right boundary that determines language of the Manchus word central axes
Suddenly it is：The boundary that the central axes leftmost side is indicated with baseline_left, it is suitable according to from left to right for longitudinal edge image Iv
Sequence scans by column, and scanning range is to arrange from the 1st column to W-wd+1, if being currently the i-th column, then 1≤i≤(W-wd+1), uses TB
The i-th column are indicated to the number of the i-th+wd-1 black pixel point arranged, then baseline_left is with corresponding to maximum TB value
Column,
The boundary of the central axes rightmost side is indicated with baseline_right, then baseline_right=baseline_
Left+wd-1,
Therefore in two-value language of the Manchus word image I, baseline_left is the left margin of the language of the Manchus word central axes,
Baseline_right is the right margin of the language of the Manchus word central axes, and wd is the width of the language of the Manchus word central axes, to position
The central axes of the language of the Manchus word out.
Claims (4)
1. the localization method that a kind of pair of language of the Manchus word correctly divides the language of the Manchus word central axes of each identification, which is characterized in that including
Following steps：
S1. word image pre-processes；
S2. the width of language of the Manchus word central axes is determined；
S3. language of the Manchus word central axes right boundary position is determined.
2. correctly divide the localization method of the language of the Manchus word central axes of each identification to language of the Manchus word as described in claim 1,
It is characterized in that, the pretreated step of language of the Manchus word image is：Language of the Manchus document is scanned, obtains the color image of language of the Manchus document, and right
The color image of language of the Manchus document carries out binaryzation, and the image of each column language of the Manchus document is extracted using vertical projection method, obtains column figure
Picture, and language of the Manchus word image is extracted using horizontal projection method to column image, place normalization then is carried out to language of the Manchus word image
The step of processing, normalized is the coordinate position for finding four points in language of the Manchus word image outermost upper and lower, left and right, with this
Four points are boundary, redefine the size of language of the Manchus word image, and the two-value after respectively indicating normalized with I, W, H is full
The height number of literary word image, the width numerical value of two-value language of the Manchus word image and two-value language of the Manchus word image.
3. correctly divide the localization method of the language of the Manchus word central axes of each identification to language of the Manchus word as described in claim 1,
The step of being characterized in that, determining language of the Manchus word axis line width be：Two-value language of the Manchus word image I is carried out using mathematical morphology
Processing, specific method be use length for 2 horizontal direction linear type morphological structuring elements, to two-value language of the Manchus word image I
Corrosion treatment is carried out, obtains the corrosion image Ie of I, and seek the longitudinal edge figure of two-value language of the Manchus word image I Yu corrosion image Ie
Picture Iv, i.e. Iv=I-Ie extract the longitudinal edge of language of the Manchus word, then carry out upright projection, system to longitudinal edge image Iv
The number of black picture element included in each column of longitudinal edge image Iv is counted, upright projection sequence is indicated with vp1, amount of orientation
Unduplicated element in vp1, and arranged according to the descending mode of numerical value from big to small, new vector v p2 is obtained, and take
Preceding 3 elements in vp2 are respectively labeled as ind1, ind2 and ind3, indicate that the width of central axes, wdmin indicate minimum with wd
Axis line width, assignment wdmin=0, wdmax indicate that maximum axis line width, T are the threshold value of maximum axis line width, assignment T
=1/3, then maximum axis line width wdmax is the value after being rounded downwards to the result of T × W, i.e.,Point
Not defining wd1, wd2, wd3 is 3 candidate axis line widths, and wd indicates the axis line width finally determined, wherein wd1=∣
∣ -1 ind1-ind2 ∣ -1, wd2=∣ ind1-ind3 ∣ -1, wd3=∣ ind2-ind3,
If (wdmin<wd1<) and (wdmin wdmax<wd2<) and (wdmin wdmax<wd3<Wdmax), then wd=
Wd1,
If (wdmin<wd1<) and (wdmin wdmax<wd2<) and (wd3 wdmax<=wdmin or wd3>=
Wdmax), then the biggish output valve as wd of wd1 and wd2 intermediate value, i.e. wd=max (wd1, wd2) are taken,
If (wdmin<wd1<) and (wdmin wdmax<wd3<) and (wd2 wdmax<=wdmin or wd2>=
Wdmax), then the biggish output valve as wd of wd1 and wd3 intermediate value, i.e. wd=max (wd1, wd3) are taken,
If (wdmin<wd2<) and (wdmin wdmax<wd3<) and (wd1 wdmax<=wdmin or wd1>=
Wdmax), then the biggish output valve as wd of wd2 and wd3 intermediate value, i.e. wd=max (wd2, wd3) are taken,
If (wdmin<wd1<) and (wd2 wdmax<=wdmin or wd2>=wdmax) and (wd3<=wdmin or
wd3>=wdmax), then wd=wd1,
If (wdmin<wd2<) and (wd1 wdmax<=wdmin or wd1>=wdmax) and (wd3<=wdmin or
wd3>=wdmax), then wd=wd2,
If (wdmin<wd3<) and (wd1 wdmax<=wdmin or wd1>=wdmax) and (wd2<=wdmin or
wd2>=wdmax), then wd=wd3,
If (wd1<=wdmin or wd1>=wdmax) and (wd2<=wdmin or wd2>=wdmax)
And (wd3<=wdmin or wd3>=wdmax), then wd=wdmax.
4. correctly divide the localization method of the language of the Manchus word central axes of each identification to language of the Manchus word as described in claim 1,
The step of being characterized in that, determining language of the Manchus word central axes right boundary position be：Indicate that central axes are most left with baseline_left
The boundary of side scans by column longitudinal edge image Iv according to sequence from left to right, and scanning range is from the 1st column to W-
Wd+1 column, if being currently the i-th column, then 1≤i≤(W-wd+1), indicates the i-th column to the i-th+wd-1 black pixel point arranged with TB
Number, then baseline_left is to indicate the central axes rightmost side with baseline_right with column corresponding to maximum TB value
Boundary, then baseline_right=baseline_left+wd-1, therefore in two-value language of the Manchus word image I,
Baseline_left is the left margin of the language of the Manchus word central axes, and baseline_right is the right side of the language of the Manchus word central axes
Boundary, wd is the width of the language of the Manchus word central axes, to orient the central axes of the language of the Manchus word.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810580666.7A CN108830270B (en) | 2015-09-29 | 2015-09-29 | Method for positioning axle wire of Manchu word for correctly segmenting each recognized Manchu word |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810580666.7A CN108830270B (en) | 2015-09-29 | 2015-09-29 | Method for positioning axle wire of Manchu word for correctly segmenting each recognized Manchu word |
| CN201510629344.3A CN105279506B (en) | 2015-09-29 | 2015-09-29 | The localization method of language of the Manchus word central axes |
Related Parent Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510629344.3A Division CN105279506B (en) | 2015-09-29 | 2015-09-29 | The localization method of language of the Manchus word central axes |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108830270A true CN108830270A (en) | 2018-11-16 |
| CN108830270B CN108830270B (en) | 2021-10-08 |
Family
ID=55148489
Family Applications (3)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510629344.3A Active CN105279506B (en) | 2015-09-29 | 2015-09-29 | The localization method of language of the Manchus word central axes |
| CN201810580288.2A Active CN108830269B (en) | 2015-09-29 | 2015-09-29 | Method for determining axial line width in Manchu words |
| CN201810580666.7A Active CN108830270B (en) | 2015-09-29 | 2015-09-29 | Method for positioning axle wire of Manchu word for correctly segmenting each recognized Manchu word |
Family Applications Before (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510629344.3A Active CN105279506B (en) | 2015-09-29 | 2015-09-29 | The localization method of language of the Manchus word central axes |
| CN201810580288.2A Active CN108830269B (en) | 2015-09-29 | 2015-09-29 | Method for determining axial line width in Manchu words |
Country Status (1)
| Country | Link |
|---|---|
| CN (3) | CN105279506B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108537229B (en) * | 2018-04-24 | 2020-06-02 | 大连民族大学 | Manchu component segmentation-based print style Manchu recognition method |
| CN108564139B (en) * | 2018-04-24 | 2021-03-19 | 大连民族大学 | Manchu component segmentation-based printed style Manchu recognition device |
| CN108596183B (en) * | 2018-04-24 | 2020-08-04 | 大连民族大学 | Over-segmentation region merging method for Manchu component segmentation |
| CN108596182B (en) * | 2018-04-24 | 2020-06-16 | 大连民族大学 | Manchu component cutting method |
| CN108564078B (en) * | 2018-04-24 | 2020-11-13 | 大连民族大学 | Method for extracting axle wire of Manchu word image |
| CN108549896B (en) * | 2018-04-24 | 2020-08-04 | 大连民族大学 | Method for deleting redundant candidate segmentation lines in Manchu component segmentation |
| CN109993236B (en) * | 2019-04-10 | 2020-11-27 | 大连民族大学 | One-shot Simese convolutional neural network-based small-sample Manchu matching method |
| CN113852730B (en) * | 2021-09-28 | 2023-12-01 | 科大讯飞股份有限公司 | Inclination angle determining method and related equipment thereof |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1834872A (en) * | 2005-03-17 | 2006-09-20 | 微软公司 | Word or character boudary-based scratch-out gesture recognition |
| CN101025791A (en) * | 2007-04-06 | 2007-08-29 | 清华大学 | Printed Monggol language text segmentation method |
| US20090169110A1 (en) * | 2005-04-20 | 2009-07-02 | Hiroaki Masuyama | Index term extraction device and document characteristic analysis device for document to be surveyed |
| CN102063619A (en) * | 2010-11-30 | 2011-05-18 | 汉王科技股份有限公司 | Character row extraction method and device |
| CN102693423A (en) * | 2012-05-15 | 2012-09-26 | 公安部第三研究所 | Method for precise positioning of license plate in strong light conditions |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5884560B2 (en) * | 2012-03-05 | 2016-03-15 | オムロン株式会社 | Image processing method for character recognition, and character recognition device and program using this method |
| CN102842034B (en) * | 2012-07-10 | 2015-09-16 | 重庆大学 | A kind of laser scanning and the device and the recognition methods that automatically identify engraving character |
| CN104298985A (en) * | 2014-08-11 | 2015-01-21 | 东莞中山大学研究院 | Iteration based image text region detecting method |
-
2015
- 2015-09-29 CN CN201510629344.3A patent/CN105279506B/en active Active
- 2015-09-29 CN CN201810580288.2A patent/CN108830269B/en active Active
- 2015-09-29 CN CN201810580666.7A patent/CN108830270B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1834872A (en) * | 2005-03-17 | 2006-09-20 | 微软公司 | Word or character boudary-based scratch-out gesture recognition |
| US20090169110A1 (en) * | 2005-04-20 | 2009-07-02 | Hiroaki Masuyama | Index term extraction device and document characteristic analysis device for document to be surveyed |
| CN101025791A (en) * | 2007-04-06 | 2007-08-29 | 清华大学 | Printed Monggol language text segmentation method |
| CN102063619A (en) * | 2010-11-30 | 2011-05-18 | 汉王科技股份有限公司 | Character row extraction method and device |
| CN102693423A (en) * | 2012-05-15 | 2012-09-26 | 公安部第三研究所 | Method for precise positioning of license plate in strong light conditions |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108830270B (en) | 2021-10-08 |
| CN105279506B (en) | 2018-07-03 |
| CN105279506A (en) | 2016-01-27 |
| CN108830269B (en) | 2021-09-07 |
| CN108830269A (en) | 2018-11-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108830270A (en) | Correctly divide the localization method of the language of the Manchus word central axes of each identification to language of the Manchus word | |
| CN108648182B (en) | Breast cancer nuclear magnetic resonance image tumor region segmentation method based on molecular subtype | |
| Song et al. | A deep learning based framework for accurate segmentation of cervical cytoplasm and nuclei | |
| CN104751142B (en) | A kind of natural scene Method for text detection based on stroke feature | |
| CN109378052B (en) | The preprocess method and system of image labeling | |
| CN109472761B (en) | Cell counting method and system based on fluorescence image | |
| CN108764229A (en) | A kind of water gauge automatic distinguishing method for image based on computer vision technique | |
| CN109002834A (en) | Fine granularity image classification method based on multi-modal characterization | |
| CN104751187A (en) | Automatic meter-reading image recognition method | |
| CN105447492B (en) | A kind of Image Description Methods based on two-dimentional local binary patterns | |
| CN105574531A (en) | Intersection point feature extraction based digital identification method | |
| CN105117740B (en) | Font identification method and apparatus | |
| CN106407979A (en) | Bill character correction method and device | |
| CN107122775A (en) | A kind of Android mobile phone identity card character identifying method of feature based matching | |
| CN105447879A (en) | Method and apparatus for detecting breast muscle in breast image | |
| Chen et al. | Method on water level ruler reading recognition based on image processing | |
| CN101916381B (en) | Object contour extraction method based on sparse representation | |
| CN106033534A (en) | Electronic paper marking method based on linear detection | |
| Zhao et al. | Feature quantification and abnormal detection on cervical squamous epithelial cells | |
| Qin et al. | Summary of scene text detection and recognition | |
| CN106407186B (en) | Establish the method and device of participle model | |
| CN109460768B (en) | Text detection and removal method for histopathology microscopic image | |
| Long et al. | Cervical cell TCT image detection and segmentation based on multi-scale feature fusion | |
| CN113989269B (en) | Traditional Chinese medicine tongue image tooth trace automatic detection method based on convolutional neural network multi-scale feature fusion | |
| CN109697287A (en) | Sentence-level bilingual alignment method and system |
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 |