CN106709892A - Rapid region expansion algorithm and device of any structure element based on stroke coding - Google Patents
Rapid region expansion algorithm and device of any structure element based on stroke coding Download PDFInfo
- Publication number
- CN106709892A CN106709892A CN201611018601.0A CN201611018601A CN106709892A CN 106709892 A CN106709892 A CN 106709892A CN 201611018601 A CN201611018601 A CN 201611018601A CN 106709892 A CN106709892 A CN 106709892A
- Authority
- CN
- China
- Prior art keywords
- interest
- area
- structural elements
- run
- length encoding
- 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.)
- Pending
Links
- 239000013598 vector Substances 0.000 claims description 19
- 238000013519 translation Methods 0.000 claims description 13
- 238000003709 image segmentation Methods 0.000 claims description 6
- 230000005484 gravity Effects 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 2
- 238000000034 method Methods 0.000 abstract description 12
- 230000014616 translation Effects 0.000 description 11
- 238000010586 diagram Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 6
- 230000000694 effects Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000010339 dilation Effects 0.000 description 1
- 230000003628 erosive effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 239000012634 fragment Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 238000011017 operating method Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
-
- 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
- G06T2207/20036—Morphological image processing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
The invention relates to a rapid region expansion algorithm and device of any structure element based on stroke coding. According to the method, stroke coding for an input image is carried out; a structure element for stroke coding is utilized to expand an area of interest after coding. The method is advantaged in that memory occupation can be effectively reduced, and the consumption time is shortened.
Description
Technical field
The present invention relates to image procossing, more particularly to the realization that area-of-interest expands.
Background technology
Morphological dilations algorithm is the Preprocessing Algorithm commonly used in image processing field, is mainly used in expanding area-of-interest
To reach connection different zones and be easy to subsequent treatment etc..In current existing area-of-interest expansion algorithm, it is mainly characterized by
Pixel is directed to be operated.Method for pixel operation can be divided into two kinds:1) each pixel uses eight tables
Show, i.e. pixel value desirable 0 to 255;2) each pixel is 0 and 1 using 1 expression, i.e. pixel value.Both pixel sides of expression
Method principle when expansive working is carried out is basically identical, and mathematic(al) representation is:Expression is tied
Constitutive element B expands to area-of-interest A, and expression formula means structural elements(It is mirror images of the structural elements B on its origin)
Translation, structural elements in image where area-of-interest AThe set of all positions being not zero with area-of-interest A common factors is i.e.
It is expansion acquired results.Wherein, method 1) ratio method 2) it is slow, because computer will be faster than to one access with the speed for calculating
Eight;In addition, method 1) EMS memory occupation amount be method 2) it is octuple.
The existing this operating method for pixel has that speed is slow, committed memory big, to the behaviour of area-of-interest
The shortcomings of work is limited by the height wide of image, is needed all pixels point in access images.That is, existing area-of-interest expansion
Time-consuming and committed memory is more for algorithm, and the pixel expansion in image boundary can be restricted.On the other hand, in actual industrial
Often required that in image is processed in real time, the time-consuming requirement to algorithm is strict, and existing this algorithm is difficult to meet
The need for practical application.
The content of the invention
The technical problem to be solved in the present invention is, for the drawbacks described above of prior art, proposes a kind of based on stroke volume
The fast area expansion algorithm of the arbitrary structures unit of code, can be effectively reduced EMS memory occupation, and shorten time-consuming.
The technical solution adopted for the present invention to solve the technical problems is:A kind of arbitrary structures based on run-length encoding are provided
The fast area expansion algorithm of unit, including:Run-length encoding is carried out to input picture;And with the structural elements of run-length encoding to coding
Area-of-interest afterwards is expanded.
In certain embodiments, it is described that run-length encoding is carried out to input picture, specifically include:It is emerging that image segmentation obtains sense
Interesting region;And to area-of-interest run-length encoding.
In certain embodiments, it is described the area-of-interest after coding is expanded with structural elements, including:To use
Structural elements expansion region of interest domain representation is translation of vector of all strokes of structural elements relative to its origin to area-of-interest
The union of result.
In certain embodiments, described selection structural elements expand to the area-of-interest after coding, specifically include:
The single stroke of structural elements is sought to the expansion results of area-of-interest;Until obtaining all strokes of the structural elements to area-of-interest
Expansion results;Union is asked to all of expansion results.
In certain embodiments, the origin of described structural elements refers to the center of gravity of structural elements.
The technical solution adopted for the present invention to solve the technical problems is still:A kind of any knot based on run-length encoding is provided
The fast area expansion gear of constitutive element, including:First module, for carrying out run-length encoding to input picture;And second module,
For being expanded to the area-of-interest after coding with the structural elements of run-length encoding.
In certain embodiments, the function that the first described module is realized is specifically included:Image segmentation obtains region of interest
Domain;And to area-of-interest run-length encoding.
In certain embodiments, the function that the second described module is realized includes:Region of interest will be expanded using structural elements
Domain representation is union of all strokes of structural elements relative to the vectorial translation result to area-of-interest of its origin.
In certain embodiments, the function that the second described module is implemented includes:Seek the single stroke pair of structural elements
The expansion results of area-of-interest;Until obtaining expansion results of all strokes of the structural elements to area-of-interest;To all of
Expansion results seek union.
The beneficial effects of the present invention are by carrying out run-length encoding to image, and using the structural elements pair of run-length encoding
The area-of-interest of image is expanded, and can be effectively reduced EMS memory occupation, and shorten time-consuming.
Brief description of the drawings
Below in conjunction with drawings and Examples, the invention will be further described, in accompanying drawing:
Fig. 1 is the schematic diagram of run-length encoding of the invention.
Fig. 2 is the flow chart of run-length encoding of the invention.
Fig. 3 is the schematic diagram of structural elements of the invention.
Fig. 4 is the schematic flow sheet of inventive algorithm.
Fig. 5 is the schematic flow sheet that the present invention carries out expansion step to area-of-interest.
Fig. 6 is stroke effect schematic diagram in structural elements shown in Fig. 3.
Fig. 7 is the result schematic diagram expanded with the structural elements shown in Fig. 3 using inventive algorithm.
Fig. 8 is the schematic block diagram of apparatus of the present invention.
Specific embodiment
In conjunction with accompanying drawing, presently preferred embodiments of the present invention is elaborated.
The present invention proposes a kind of fast area expansion algorithm of the arbitrary structures unit based on run-length encoding, and it is carried out to image
Run-length encoding, and structural elements using run-length encoding expand to the area-of-interest of image.Although not being directed to pixel
The expansion algorithm of operation is directly perceived, but its speed and EMS memory occupation are but better than the algorithm for pixel operation, and departing from
Image, for example:It is wide, high, the limitation with region shape.The applicable technical field of inventive algorithm includes, but are not limited to:At image
Reason Blob extends area-of-interest when analyzing;Opening and closing operations are combined into erosion algorithm and seek region geometry characteristic parameter etc..
Inventive algorithm is generally comprised:Run-length encoding is carried out to input picture;And with the structural elements of run-length encoding to compiling
Area-of-interest after code is expanded.
On carrying out run-length encoding to input picture.The pixel of different gray values is combined and constitutes image, by
Cause that image shows different patterns in the difference of pixel gray value, often need to split image in image procossing with
Obtain region interested.Area-of-interest is typically directly represented with pixel in existing algorithm, for example:Increase income storehouse OpenCV's
ROI, pixel represents that the drawback that area-of-interest is present is big EMS memory occupation.For with the gray-scale map of 8 bit depth, if one
Piece area-of-interest contains 300,000 pixels, then its EMS memory occupation amount MemoryUsage=300000bytes, about 292KB,
According to run-length encoding mode, EMS memory occupation amount may as little as 3KB, the two difference is nearly 100 times, it is seen that run-length encoding image is to subtracting
The occupancy of few program resource is effective.
Run-length encoding can be divided into two kinds of every trade journey and row stroke, and every trade journey structure rises comprising line number Row, stroke where stroke
Begin row Cb and stroke end row Ce;Row stroke includes row number Col, stroke initial row Rb and stroke end row Re where stroke.
Run-length encoding of the invention is realized based on every trade journey.Continuous image vegetarian refreshments composition in every a line of area-of-interest
One section of region be referred to as a stroke.Uniquely one stroke of sign need to provide line number, starting row and end column.Bag in one stroke
The points of the pixel for containing are not limited.If each component of stroke accounts for 6 words with two byte representations, a stroke
Section.It is made up of limited stroke per a piece of area-of-interest.It is likely to comprising several strokes per a line.Referring to Fig. 1, Fig. 1 is
The schematic diagram of run-length encoding of the invention.Enumerated as one, Fig. 1 shows a pixel 101, two strokes 103,105 with
And area-of-interest 107.
Referring to Fig. 2, Fig. 2 is the flow chart of run-length encoding of the invention.The run-length encoding process of area-of-interest is specifically wrapped
Include following steps:
S201, image segmentation obtain area-of-interest.That is, region of interest is obtained by image partition methods such as Threshold segmentations
Domain.
S202, application internal memory.That is, a piece of internal memory is opened up in computing device to be used to preserve stroke.
S203, to area-of-interest run-length encoding.That is, taken successively since area-of-interest minimum row minimum row pixel
Line number Row, starting row Cb and end column Ce, obtain first stroke and are stored in the internal memory opened up;Line by line will be whole
Encoding region of interest is represented for stroke.
S204, whether enough judge internal memory that abovementioned steps S202 opens up, it is enough if, go to step S206, not enough, turn
Step S205.
During S205, again application internal memory then need to open up again a piece of big internal memory and legacy data are replicated into so far section internal memory,
Discharge original internal memory.
S206, coding complete to terminate coding.
By after above-mentioned steps, area-of-interest is encoded as stroke representation, and all strokes according to row order
Preserve successively, often all strokes of row are preserved in strict accordance with row order again.
On being expanded to the area-of-interest after coding with the structural elements of run-length encoding.Referring to Fig. 3, Fig. 3 is this hair
The schematic diagram of bright structural elements.Structural elements need the origin 301 of definition structure unit, expansion algorithm to use this structure when choosing
The pixel of unit is relative to the result after the mirror image of origin 301.If structural elements are B, the result after mirror image isArea-of-interest is
A.Intuitivism apprehension expansion refers to the translation structure unit in whole image regionAll structural elements origins (i.e. weight of structural elements
The heart) position be contained in area-of-interest A composition set as expansion results, mathematic(al) representation be the first formula:If image where area-of-interest A contains m pixel, structural elements B contains n pixel,
Then the complexity of algorithm is Ο (mn).
The present invention dexterously does linear transformation herein:Structural elements are also denoted as using structural elements B expansions area-of-interest ARelative to the union of the vectorial translation result to area-of-interest A of its origin, mathematic(al) representation is second public to all pixels point
Formula:Wherein, vectorAs indicated by the arrows in fig. 3.
If pixel is substituted with stroke, all pixels point translation vector of each stroke 302Can be analyzed to a row to
AmountWith two column vectorsSet.Because in per a lineThere is coincidence, such as vector in vectorWithOverlapPart, their coefficient results and maximum vectorIt is single
Solely the result of effect is identical, therefore for certain one stroke 302, need to only calculate the starting point and terminating point vector of stroke 302
To the union of area-of-interest A translations.It follows that containing how many pixels regardless of certain one stroke, it is finally only needed
Influence of two pixels to expansion algorithm is calculated, therefore the second formula is dropped compared to the algorithm complex of the first formula
It is low, it is especially prominent in the larger situations of structural elements B.
Further, since area-of-interest A is also for stroke is represented, one amount of calculation of stroke of translation is equivalent to translation 1.5
Pixel amount of calculation.But a stroke can represent hundreds and thousands of pixels, therefore the algorithm complex represented by the first formula
Further reduced.By taking Rectangle structure cell as an example, it is assumed that:The line number of structural elements B is RB, columns be CB, common n pixel;Sense
The line number of interest region A is RA, columns be CA, common m pixel;Then the complexity of the first formula is Ο (mn), public using second
The complexity of formula is Ο (3RARB), usual 3RARBMuch smaller than mn, it can be seen that, inventive algorithm complexity is substantially reduced, the time
Consumption can reduce a lot.
Referring to Fig. 4, Fig. 4 is the schematic flow sheet of inventive algorithm.Inventive algorithm generally comprises following steps:S401、
Image segmentation is carried out to target image, area-of-interest is obtained.First, an appropriate threshold value is chosen, target image is then traveled through
Middle all pixels point, meet threshold requirement is then area-of-interest pixel, finally by the pixel of all area-of-interests
Mark and preserve and form area-of-interest.
S402, run-length encoding is carried out to area-of-interest.It is first stroke with the minimum row of area-of-interest minimum row
Starting point, extracts continuous pixel fragment for a stroke successively, until the coding of whole area-of-interest is completed, coding step
Suddenly referring to Fig. 1 and Fig. 2.
S403, area-of-interest is expanded with structural elements.The need for inventive algorithm can be according to practical application, choose
Arbitrary structures unit expands to area-of-interest.If choosing square structure unit, it is clear that expansion results will become corner angle;If choosing
Circular configuration unit, then can make area-of-interest corner become mellow and full.
Referring to Fig. 5, Fig. 5 is the schematic flow sheet that the present invention carries out expansion step to area-of-interest.Expansion of the invention
Process is specifically included:
S501, the single stroke of structural elements is sought to the expansion results of area-of-interest.Specifically have, each to structural elements B
Stroke enters row vector decomposition, obtains line direction vector and two maximum column direction vectors, a row vector as shown in Figure 3With
Two column vectorsWherein, row vectorDetermine the row Row of all strokes of area-of-interest A, two arrange to
AmountDetermine the starting row Cb and end column Ce of all strokes of area-of-interest A.Referring to Fig. 6, Fig. 6 is Fig. 3 institutes
Show stroke effect schematic diagram in structural elements.Wherein, under background 600, original area is designated 601, column vectorMake sense emerging
What interesting region A moved to left rear region is designated 603, column vectorMake what area-of-interest A moved to right rear region to be designated 602, translation
What union afterwards was the column direction expansion results of single stroke is designated 604.New starting row Cb and end column Ce is such as Fig. 6
Shown in middle union part 604.With reference to row vectorEffect to area-of-interest A, you can the single stroke for obtaining structural elements B is swollen
Swollen result.
The mathematic(al) representation of above-mentioned single stroke expansion is:
Cb'=Cb+vcl;Ce'=Ce+vcr;Row'=Row+vr.
S502, above-mentioned steps S501 is repeated, until obtaining expansion of all strokes of structural elements B to area-of-interest A
As a result.
S503, all results that step S502 is obtained are sought into union, as final expansion results.Referring to Fig. 7, Fig. 7 is to adopt
The result schematic diagram expanded with the structural elements shown in Fig. 3 with inventive algorithm.Used as an example, Fig. 7 is illustrated
Result 702 after original area 701 and expansion under background 700.
Referring to Fig. 8, Fig. 8 is the schematic block diagram of apparatus of the present invention.Enumerated as in one kind, the device 800 includes:First
Module 801, for carrying out run-length encoding to input picture;And second module 802, for the structural elements of run-length encoding to compiling
Area-of-interest after code is expanded.
The function that first module 801 is realized is specifically included:Image segmentation obtains area-of-interest;And to region of interest
Domain run-length encoding.
The function that second module 802 is realized includes:It is structural elements institute that structural elements will be used to expand region of interest domain representation
There is stroke relative to the union of the vectorial translation result to area-of-interest of its origin.Specifically, second module 802 has
The function that body is realized includes:The single stroke of structural elements is sought to the expansion results of area-of-interest;Until obtaining the structural elements institute
There are expansion results of the stroke to area-of-interest;Union is asked to all of expansion results.
It is understood that above-mentioned two modules 801 and 802 both can correspond to the two of the program run on processor
Individual software module, it is also possible to corresponding to special two hardware modules for building.
Beneficial effects of the present invention include:Structural elements are with area-of-interest using after run-length encoding, and algorithm EMS memory occupation is few,
It is particularly well-suited to the embedded platform being had certain limitations to resource;Image represents that the least unit of image is row after run-length encoding
Journey, the pixel number that each stroke can be included is unrestricted, therefore the number of strokes of area-of-interest is necessarily fewer than pixel number, with
The algorithm for traveling through pixel then becomes simple compared to traversal stroke, and taking will also be reduced;In addition, after using run-length encoding, to sense
The operation in interest region directly preserves the region of interest of arbitrary shape arbitrary size departing from the width of image and the limitation of height
Domain, no data redundancy, applicability is wider.
It should be appreciated that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations, to ability
For field technique personnel, the technical scheme described in above-described embodiment can be modified, or it is special to which part technology
Levying carries out equivalent;And these modifications and replacement, should all belong to the protection domain of appended claims of the present invention.
Claims (10)
1. it is a kind of based on run-length encoding arbitrary structures unit fast area expansion algorithm, it is characterised in that including:Input is schemed
As carrying out run-length encoding;And the area-of-interest after coding is expanded with the structural elements of run-length encoding.
2. algorithm according to claim 1, it is characterised in that:Described carries out run-length encoding to input picture, specific bag
Include:Image segmentation obtains area-of-interest;And to area-of-interest run-length encoding.
3. algorithm according to claim 1, it is characterised in that:Described is entered with structural elements to the area-of-interest after coding
Row expansion, including:Structural elements will be used to expand vector of the region of interest domain representation for all strokes of structural elements relative to its origin
To the union of the translation result of area-of-interest.
4. algorithm according to claim 3, it is characterised in that:Described selection structural elements are to the area-of-interest after coding
Expanded, specifically included:The single stroke of structural elements is sought to the expansion results of area-of-interest;Until obtaining the structural elements institute
There are expansion results of the stroke to area-of-interest;Union is asked to all of expansion results.
5. algorithm according to claim 3, it is characterised in that:The origin of described structural elements refers to the center of gravity of structural elements.
6. it is a kind of based on run-length encoding arbitrary structures unit fast area expansion gear, it is characterised in that:Including:First mould
Block, for carrying out run-length encoding to input picture;And second module, for the structural elements of run-length encoding to the sense after coding
Interest region is expanded.
7. device according to claim 6, it is characterised in that:The function that the first described module is realized is specifically included:Figure
As segmentation obtains area-of-interest;And to area-of-interest run-length encoding.
8. device according to claim 6, it is characterised in that:The function that the second described module is realized includes:To use
Structural elements expansion region of interest domain representation is translation of vector of all strokes of structural elements relative to its origin to area-of-interest
The union of result.
9. device according to claim 8, it is characterised in that:The function that the second described module is implemented includes:Ask
Expansion results of the single stroke of structural elements to area-of-interest;Until obtaining all strokes of the structural elements to area-of-interest
Expansion results;Union is asked to all of expansion results.
10. device according to claim 6, it is characterised in that:The origin of described structural elements refers to the center of gravity of structural elements.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611018601.0A CN106709892A (en) | 2016-11-15 | 2016-11-15 | Rapid region expansion algorithm and device of any structure element based on stroke coding |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611018601.0A CN106709892A (en) | 2016-11-15 | 2016-11-15 | Rapid region expansion algorithm and device of any structure element based on stroke coding |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106709892A true CN106709892A (en) | 2017-05-24 |
Family
ID=58940589
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611018601.0A Pending CN106709892A (en) | 2016-11-15 | 2016-11-15 | Rapid region expansion algorithm and device of any structure element based on stroke coding |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106709892A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106846424A (en) * | 2016-12-30 | 2017-06-13 | 昂纳自动化技术(深圳)有限公司 | Fast area expansion algorithm and device based on run-length encoding |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104050473A (en) * | 2014-05-20 | 2014-09-17 | 中国人民解放军理工大学 | Road data extraction method based on rectangular neighborhood analysis |
CN105809720A (en) * | 2015-01-20 | 2016-07-27 | 台湾积体电路制造股份有限公司 | System and method for maskless direct write lithography |
CN105931198A (en) * | 2016-04-14 | 2016-09-07 | 西安工程大学 | Icing insulator image enhancement method based on wavelet transformation |
CN106023076A (en) * | 2016-05-11 | 2016-10-12 | 北京交通大学 | Splicing method for panoramic graph and method for detecting defect state of guard railing of high-speed railway |
-
2016
- 2016-11-15 CN CN201611018601.0A patent/CN106709892A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104050473A (en) * | 2014-05-20 | 2014-09-17 | 中国人民解放军理工大学 | Road data extraction method based on rectangular neighborhood analysis |
CN105809720A (en) * | 2015-01-20 | 2016-07-27 | 台湾积体电路制造股份有限公司 | System and method for maskless direct write lithography |
CN105931198A (en) * | 2016-04-14 | 2016-09-07 | 西安工程大学 | Icing insulator image enhancement method based on wavelet transformation |
CN106023076A (en) * | 2016-05-11 | 2016-10-12 | 北京交通大学 | Splicing method for panoramic graph and method for detecting defect state of guard railing of high-speed railway |
Non-Patent Citations (2)
Title |
---|
康飞: ""基于机器视觉的轨道检测***研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
郭云彪等: ""二值图像信息隐藏分析"", 《天津大学学报》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106846424A (en) * | 2016-12-30 | 2017-06-13 | 昂纳自动化技术(深圳)有限公司 | Fast area expansion algorithm and device based on run-length encoding |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11908244B2 (en) | Human posture detection utilizing posture reference maps | |
KR102640237B1 (en) | Image processing methods, apparatus, electronic devices, and computer-readable storage media | |
CN110443842B (en) | Depth map prediction method based on visual angle fusion | |
CN110738207B (en) | Character detection method for fusing character area edge information in character image | |
CN109711413A (en) | Image, semantic dividing method based on deep learning | |
CN108416266B (en) | Method for rapidly identifying video behaviors by extracting moving object through optical flow | |
CN109903292A (en) | A kind of three-dimensional image segmentation method and system based on full convolutional neural networks | |
CN110689599B (en) | 3D visual saliency prediction method based on non-local enhancement generation countermeasure network | |
CN111723585A (en) | Style-controllable image text real-time translation and conversion method | |
CN107784301A (en) | Method and apparatus for identifying character area in image | |
CN105678813A (en) | Skin color detection method and device | |
CN111275034B (en) | Method, device, equipment and storage medium for extracting text region from image | |
CN104952083B (en) | A kind of saliency detection method based on the modeling of conspicuousness target background | |
CN100375124C (en) | A skeletonized object rebuild method | |
CN110399760A (en) | A kind of batch two dimensional code localization method, device, electronic equipment and storage medium | |
CN112884758A (en) | Defective insulator sample generation method and system based on style migration method | |
CN111414938A (en) | Target detection method for bubbles in plate heat exchanger | |
CN113177526A (en) | Image processing method, device and equipment based on face recognition and storage medium | |
CN106709892A (en) | Rapid region expansion algorithm and device of any structure element based on stroke coding | |
CN107832732A (en) | Method for detecting lane lines based on ternary tree traversal | |
CN114862716B (en) | Image enhancement method, device, equipment and storage medium for face image | |
CN112861860B (en) | Text detection method in natural scene based on upper and lower boundary extraction | |
CN115861349A (en) | Color image edge extraction method based on reduction concept structural elements and matrix sequence | |
Alexeev et al. | Using a fully connected convolutional network to detect objects in images | |
CN106600655A (en) | Rapid area erosion algorithm and device based on random structural element of run-length encoding |
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 | ||
TA01 | Transfer of patent application right | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20200611 Address after: 518000 floor 1, building 604, building 3, workshop 1, auna group, Zhukeng community, Longtian street, Pingshan District, Shenzhen City, Guangdong Province Applicant after: Ona Industrial Technology (Shenzhen) Co., Ltd Address before: Cui Jing Road Shenzhen City, Guangdong province 518000 Pingshan New District No. 35 Applicant before: O-NET AUTOMATION TECHNOLOGY (SHENZHEN) Co.,Ltd. |
|
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170524 |