CN107038708A - Application of the image recognition algorithm in paper-cut effect - Google Patents
Application of the image recognition algorithm in paper-cut effect Download PDFInfo
- Publication number
- CN107038708A CN107038708A CN201710265951.5A CN201710265951A CN107038708A CN 107038708 A CN107038708 A CN 107038708A CN 201710265951 A CN201710265951 A CN 201710265951A CN 107038708 A CN107038708 A CN 107038708A
- Authority
- CN
- China
- Prior art keywords
- image
- edge
- paper
- cut
- effect
- 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
- 230000000694 effects Effects 0.000 title claims abstract description 34
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 32
- 238000001514 detection method Methods 0.000 claims abstract description 29
- 238000000034 method Methods 0.000 claims abstract description 23
- 238000003708 edge detection Methods 0.000 claims abstract description 10
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 238000012805 post-processing Methods 0.000 claims abstract description 5
- 230000008569 process Effects 0.000 claims description 9
- 238000001914 filtration Methods 0.000 claims description 8
- 230000009467 reduction Effects 0.000 claims description 5
- 238000007689 inspection Methods 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 abstract description 9
- 238000000605 extraction Methods 0.000 abstract description 5
- 238000003709 image segmentation Methods 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 8
- 230000008859 change Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000014509 gene expression Effects 0.000 description 3
- 238000009499 grossing Methods 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 238000009826 distribution Methods 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 238000005303 weighing Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010421 pencil drawing Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 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 by the use of local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/001—Texturing; Colouring; Generation of texture or colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/20—Drawing from basic elements, e.g. lines or circles
- G06T11/203—Drawing of straight lines or curves
-
- 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 by the use of local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G06T5/70—
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
Application of the public image recognizer of the present invention in paper-cut effect, including carry out image preprocessing, realize rim detection, generate sketch map, realize paper-cut sun carve, stamp effect and lines post-processing, image recognition and extraction algorithm are applied in traditional culture paper-cut first, paper-cut is transferred on line under line by technological meanses, the image processing algorithm usually studied is incorporated and lived, the edge detection algorithm of image segmentation is employed simultaneously, with five kinds of operator overlap-add procedures, the higher target image of precision can be extracted.
Description
Technical field：
The invention belongs to image recognition and extractive technique field, and in particular to application of the image recognition algorithm in paper-cut effect.
Background technology：
Paper-cut is that one kind cuts texturing line with scissors or icking tool on paper, for decorateeing life or coordinating the among the people of other folk activities
Art.In China, paper-cut has extensive mass foundation, blends in the social life of people of all nationalities, is various folk activities
Important component., more can not be by paper-cut and image recognition and prior art can not easily process the image into paper-cut effect
Algorithm is combined.
The present invention first applies image recognition and extraction algorithm in traditional culture paper-cut, by technological meanses by paper-cut
It is transferred under from line on line, is incorporated and lived with the image processing algorithm usually studied, the rim detection for employing image segmentation is calculated
Method, with five kinds of operator overlap-add procedures, can extract the higher target image of precision.
The content of the invention：
In view of the above-mentioned problems, comprising the following steps the invention provides image recognition algorithm in the application of paper-cut effect：
A. image preprocessing is carried out：To artwork noise-removed filtering, cromogram is switched into gray-scale map；
B. rim detection is realized：The edge detection algorithm split using image carries out edge inspection by differential operator to gray-scale map
Survey；
C. sketch map is generated：Numeral in the image array for carrying out rim detection is replaced by the numeral representated by grey, obtained
Sketch colored line, generates sketch map；
D. paper-cut sun quarter, stamp effect are realized：Sketch map to generation is coloured, and it is that edge is red that sun, which carves effect, and background is white
Color, stamp effect is edge white, and background is red；
E. lines post-processing：According to edge mean breadth, target image is corroded or expansion process, obtain smoother
Profile.
It is preferred that, the step（1）In be, by Gaussian filter filtering and noise reduction, cromogram to be entered to artwork noise-removed filtering
Row binary conversion treatment, is converted to gray level image.
It is preferred that, the step（2）In rim detection be utilize the difference of object and background in certain picture characteristics
To realize detection, the difference includes gray scale, color or textural characteristics.
Beneficial effect of the present invention：The present invention is by carrying out image preprocessing, realizing rim detection, generation sketch map, realization
Paper-cut sun quarter, stamp effect and lines post-processing, first apply image recognition and extraction algorithm in traditional culture paper-cut,
Paper-cut is transferred on line under line by technological meanses, the image processing algorithm usually studied is incorporated and lived, while using
The edge detection algorithm of image segmentation, with five kinds of operator overlap-add procedures, can extract the higher target image of precision.
Brief description of the drawings：
Fig. 1 is the gray level image obtained after image preprocessing.
Fig. 2 is the Matlab analogous diagrams that Roberts operators obtained after edge treated.
Fig. 3 is the Matlab analogous diagrams that Sobel operators obtained after edge treated.
Fig. 4 is the Matlab analogous diagrams that Prewitt operators obtained after edge treated.
Fig. 5 is the Matlab analogous diagrams that Canny operators obtained after edge treated.
Embodiment：
To make the object, technical solutions and advantages of the present invention of greater clarity, below by the specific implementation shown in accompanying drawing
Example describes the present invention.However, it should be understood that these descriptions are merely illustrative, and it is not intended to limit the scope of the present invention.This
Outside, in the following description, the description to known features and technology is eliminated, to avoid unnecessarily obscuring idea of the invention.
The image recognition algorithm that the present invention is provided comprises the following steps in the application of paper-cut effect：
A. image preprocessing is carried out：First input picture artwork is pre-processed, is that algorithm calculating speed reaches with effect below
It is optimal to prepare.Gaussian filter filtering and noise reduction is used to image artwork, cromogram is subjected to binary conversion treatment, gray scale is converted to
Image, such as Fig. 1.
B. rim detection is realized：Rim detection occupies special position in image procossing and computer vision, and it is bottom
One of most important link in layer visual processes, is also the basis for realizing the image segmentation based on border.In the picture, border table
The termination of a bright characteristic area and the beginning of another characteristic area, the internal feature or attribute of border institute separation region are one
Cause, and the feature or attribute inside different zones are different, the detection at edge exactly utilizes object and background in certain figure
Realized as the difference in characteristic.This species diversity includes gray scale, color or textural characteristics.Rim detection is actually inspection
The position that altimetric image characteristic changes.Functional derivative reflects the significance degree of variation of image grayscale, the local pole of first derivative
The zero crossing of big value and second dervative is all that variation of image grayscale is greatly local.Therefore these derivative values can be regard as respective point
Boundary intensity, the method by setting thresholding, extract border point set.
Rim detection based on first derivative, gradient is the first derivative of image correspondence two-dimensional function：
The amplitude of gradient can be weighed with following three kinds of norms:
Gradient direction is function maximum rate of change direction:
Conventional edge detection operator has Roberts operators, Sobel operators, Prewitt operators, LoG operators and Canny operators,
Because image is made up of discrete pixel, these operators will use difference approximation partial derivative.
It is four kinds of operator edge detections based on first derivative below：
1) Roberts crossover operators
Roberts edge detection operators are at a kind of operator that edge is found using local difference operator, Robert operator images
Result edge is not very smooth after reason.Through analysis, because Robert operators would generally be produced in the region near image border
Wider response, therefore the edge image for using above-mentioned operator to detect need to often do micronization processes, the precision of edge positioning is not very high.
Two convolution kernels of Roberts operators are respectively
The amplitude of gradient is weighed using 1 norm：
Roberts operators are preferable to the image effect with precipitous low noise.
The Matlab analogous diagrams such as Fig. 2 obtained after the processing of Roberts operators.
2) Sobel operators
Sobel Operator（Sobel operator）It is one of operator in image procossing, is mainly used as rim detection.In technology
On, it is a discreteness difference operator, for the approximation of the gradient of computing brightness of image function.Make in any point of image
Use this operator, it will produce corresponding gradient vector or its law vector.
Two convolutional calculation cores of Sobel operators are respectively：
The amplitude of gradient is weighed using ∞ norms。
The Sobel operators image procossing more to gray scale gradual change and noise is obtained preferably.
The Matlab analogous diagrams such as Fig. 3 obtained after the processing of Sobel operators.
In Edge check, a kind of conventional template is Sobel operators.Sobel operators have two, and one is detection water
Flat edge；Another is the vertical flat edge of detection.Compared with Prewitt operators, Sobel operators are for the position of pixel
Influence is weighted, and can reduce edge blurry degree, therefore effect is more preferable.
Another form of Sobel operators is isotropism Sobel (Isotropic Sobel) operator, also there is two, one
It is detection level edge, another is the vertical flat edge of detection.Isotropism Sobel operators and common Sobel operators
Compare, its position weight coefficient is more accurate, when detecting the edge of different directions, the amplitude of gradient is consistent.Due to building
The particularity of image, we it can be found that processing the type image outline when, and need not to gradient direction carry out computing, institute
The processing method of isotropism Sobel operators is not provided with program.
Because Sobel operators are the forms of filter operator, for extracting edge, it is possible to use fast convolution function, simply
Effectively, therefore it is widely used.Only drawback is that, Sobel operators do not distinguish the main body of image strictly with background
Come, be in other words exactly that Sobel operators are not based on gradation of image and handled, because Sobel operators do not have strict simulation people
Vision physiological feature, so extract image outline it is sometimes not satisfactory.When piece image is observed, we
Often it is first noted that be the image part different from background, exactly this part highlights main body, based on the theory, I
Give following thresholding contours extract algorithm, the algorithm has mathematically proved required when pixel meets normal distribution
Solution is optimal.
3) Prewitt operators
Prewitt operators are a kind of rim detections of first order differential operator, using above and below pixel, the gray scale difference of left and right adjoint point,
Extremum extracting edge is reached in edge, removes part pseudo-edge, there is smoothing effect to noise.Its principle is in image space
Neighborhood convolution is carried out using both direction template and image to complete, one detection level edge of the two direction templates, one
Individual detection vertical edge.
Two convolutional calculation cores of Prewitt operators are respectively
As Sobel operators, output is used as using ∞ norms.
The Prewitt operators image procossing more to gray scale gradual change and noise is obtained preferably.
The Matlab analogous diagrams such as Fig. 4 obtained after the processing of Prewitt operators.
4) Canny operators
The gradient of Canny operators is the derivative calculations with Gaussian filter, and the method at detection edge is to find image gradient
Local maximum.Canny methods detect strong edge and weak edge using two threshold values respectively, and and if only if weak edge and strong side
When edge is connected, weak edge just can be comprising in the output.Therefore the method is not susceptible to the interference of noise, is able to detect that weak side
Edge.
Canny algorithm steps：
A) Gaussian filter smoothed image is used；
B) amplitude and the direction of filtered image gradient are calculated；
C) to gradient magnitude application non-maxima suppression, its process is to find out the Local modulus maxima in image gradient, other
Edge of the non local maximum point zero to be refined；
D) with the detection of dual threashold value-based algorithm and connection edge, two threshold values T1 and T2 are used（T1>T2）, T1 be used for find every line
Section, T2 is used for the extension in the both direction of these line segments and finds the breaking part at edge, and connects these edges.
Canny key theory：
Noise reduction:Any edge detection algorithm is impossible to work well in undressed initial data, so the first step
That convolution is made to initial data and Gauss mask, obtained image with original image compared with some slight obscure（blurred）.
So, a single pixel noise becomes to have little to no effect on the image by Gaussian smoothing.
Find the brightness step in image:Edge in image may point to different directions, so Canny algorithms
Use 4 mask detection levels, the edge of vertical and diagonal.The convolution that original image and each mask are made
Store.Put us for each and identify the direction at the edge of maximum on this aspect and generation.So we
Just from original image generate image in each point brightness gradient map and brightness step direction.
Following limb in the picture:Higher brightness step is relatively likely to be edge, but the definite value of neither one
It is that edge is much and be not to limit great brightness step, so Canny has used hysteresis threshold.
Hysteresis threshold needs two threshold value-high thresholds and Low threshold.Assuming that the important edges in image are all continuous songs
Line, so we can just track the part obscured in given curve, and avoid working as the noise pixel without constituent curve
Into edge.So we are since a larger threshold value, this will be identified before the true edge that we relatively firmly believe, use
Derived directional information, we track whole edge in the picture since these real edges.When tracking,
We thus until us can return to starting point using less threshold value with the blurred portions of aircraft pursuit course.Once this
Process is completed, and we have just obtained a bianry image, and every indicates whether it is a marginal point.One acquisition sub-pixel precision
It is the zero crossing that Second order directional is detected in gradient direction that the improvement at edge, which is realized,.
It meets sign condition in three rank directional derivatives of gradient direction
The metric space that wherein Lx, Ly ... Lyyy expressions are obtained with the smooth original image of Gaussian kernel represents that L is calculated and obtained
Partial derivative.The edge segments obtained in this way are full curves, and so there is no need to the improvement of other Edge track.It is stagnant
Threshold value can be used for sub-pixel edge detection afterwards.
Canny parameters：
Canny algorithms include many adjustable parameters, and they will have influence on time and the actual effect of the calculating of algorithm.Gauss is filtered
The size of ripple device：All smoothing filters of the first step will directly affect the result of Canny algorithms.Less wave filter production
Raw blur effect is also less, can thus detect smaller, the obvious fine rule of change.The fuzzy effect that larger wave filter is produced
Fruit is also more, and one piece of larger image-region is painted to the color value of a specified point.The result so brought is exactly for inspection
Survey larger, smooth edge more useful, the edge of such as rainbow.Threshold value：A threshold value is used using two threshold value ratios more
Flexibly, but its common problem still with the presence of threshold value.The threshold value of setting is too high, may miss important information；Threshold value mistake
It is low, it will minor matters information is taken seriously will.It is difficult to provide a generic threshold value for being applied to all images.There is presently no one
The individual implementation method by checking.
Canny operator core formula：
The Gaussian functions are taken to be
Canny operators are set up in two-dimensional convolutionOn basis, edge strength and direction are obtained, passes through threshold value
To detect edge.
WillTwo-dimensional convolution template decomposition be two one-dimensional filtering devices, obtain
In formula
It can be seen that
Then the two templates are carried out convolution with f (x, y) respectively, obtained
Order
Then A (i, j) reflecting edge intensity, a (i, j) is perpendicular to the direction at edge.
The Matlab analogous diagrams such as Fig. 5 obtained after the processing of Canny operators.
5) Laplacian of Gaussian algorithms（Laplace operator, abbreviation LoG operator definitions）
LoG operator definitions：
For scalar field function f, for a scalar of the divergence of the scalar field gradient, i.e., for vector field function, f is the vector
The gradient of field divergence subtracts a vector of the vector field curl, i.e. Laplace operator.
Laplace operator is a Second Order Differential Operator in n dimensions Euclidean space, is defined as gradient（▽f）Dissipate
Degree（▽·f）.So if f is the real function that second order can be micro-, then f Laplace operator is defined as：
F Laplace operator is also all non-mixed second-order partial differential coefficients in Cartesian coordinates xi：
As a Second Order Differential Operator, C function is mapped to C function by Laplace operator, for k >=2.Expression formula (1)（Or
(2)）Define an operator Δ:C (R) → C (R), or more generally, define an operator Δ:C (Ω) → C (Ω) is right
In any opener Ω.
The Laplace operator of function is also the mark of the Hessian matrix of the function：
LoG operators are promoted：
Laplace operator can be generalized to non-euclidean space with certain method, and at this moment it is possible to be that ellipse is calculated
Son, hyperbolic operator, or ultra-hyperbolic type operator.
In minkowskian space, Laplace operator is changed into reaching bright Bel's operator：
Up to bright Bel's operator commonly used to expression Klein-Gordon equation and four-dimensional wave equation.Symbol before 4th item
Number it is negative sign, and is then positive sign in Euclidean space.Factor c is desirable, because time and space be not generally with
With unit weigh；If x directions are with cun weighing, y directions are with centimetre weighing, it is also desirable to similar factor.
Laplce-Marco Beltrami operator
Main entry：Laplce-Marco Beltrami operator
Laplace operator can also be extended to the Elliptic Operator being defined in Riemann manifold, and referred to as Laplce-Bel is special
Rummy operator.The hyperbolic operator in pseudo-Riemannian manifold is then extended to up to bright Bel's operator.Laplce-Marco Beltrami operator
The operator run in tensor field can also be extended to（Also referred to as Laplce-Marco Beltrami operator）.
The method that another Laplace operator is generalized to pseudo-Riemannian manifold, is calculated by Laplce-De Lamu
Son, it runs on differential form.This just can pass through Weitzenb＆ouml；Ck identities come and Laplce-Marco Beltrami
Operator is connected.
It can be seen that Canny methods detect strong edge and weak edge respectively using two threshold values by above-mentioned four kinds of operators, this
Method is not susceptible to the interference of noise, is able to detect that weak edge, is achieved between noise suppressed and rim detection preferably
Balance, edge is most clear, and noise is minimum, more preferable than the effect that Roberts operator, Prewitts operators, Sobel operators are obtained.
C. sketch map is generated：0 and 1 in the picture matrix for carrying out rim detection is subjected to reversing of position, new figure is obtained
Numeral in obtained new picture matrix, is changed into the numeral representated by grey by piece, and resulting new picture is white background
The pencil drawing effect of grey lines.Representative numeral is grey from shallow to deep：238、229、220、210、201、191、181、
170th, 160, the numeral in original picture matrix is replaced by the numeral representated by grey, sketch colored line, generation element is obtained
Tracing.
D. paper-cut sun quarter, stamp effect are realized：Dynamic finds threshold value, inverse, colour changing is carried out to sketch map, it is edge that sun, which is carved,
Red, background white, stamps then opposite.
E. lines post-processing：In the step of contour line extraction, target image is corroded or expanded due to needing
Operation, will also do closed operation to the edge image of binaryzation, the distribution situation at this meeting subtle effects gradation of image and edge, institute
The contour line key point extracted is possible to not on the contour line of original image, and has certain deviation with contour line.Cause
This after edge, sun quarter, stamp effect process is extracted, can there is prominent portion in image outline lines according to edge mean breadth
Divide or fall in this edge noise suddenly, overstriking is carried out to lines（Expansion process）, narrow（Corrosion treatment）, relatively put down
Sliding profile.
After the present invention splits target image, edge extracting is carried out, the effect of extraction and the degree of accuracy greatly have impact on most
The reduction degree of paper-cut afterwards, the present invention has attempted existing edge detection algorithm, simulate Roberts operators, Prewitt operators,
Sobel operators, the design sketch of four kinds of algorithms of different processing of Canny operators, first three species diversity less, imitate by the 4th kind of Canny operator
Fruit is more preferably.
The general principle and principal character and advantages of the present invention of the present invention has been shown and described above.The technology of the industry
Personnel are it should be appreciated that the present invention is not limited to the above embodiments, and the simply explanation described in above-described embodiment and specification is originally
The principle of invention, without departing from the spirit and scope of the present invention, various changes and modifications of the present invention are possible, these changes
Change and improvement all fall within the protetion scope of the claimed invention.The claimed scope of the invention by appended claims and its
Equivalent thereof.
Claims (3)
1. image recognition algorithm is in the application of paper-cut effect, it is characterised in that：Comprise the following steps：
A. image preprocessing is carried out：To artwork noise-removed filtering, cromogram is switched into gray-scale map；
B. rim detection is realized：The edge detection algorithm split using image carries out edge inspection by differential operator to gray-scale map
Survey；
C. sketch map is generated：Numeral in the image array for carrying out rim detection is replaced by the numeral representated by grey, obtained
Sketch colored line, generates sketch map；
D. paper-cut sun quarter, stamp effect are realized：Sketch map to generation is coloured, and it is that edge is red that sun, which carves effect, and background is white
Color, stamp effect is edge white, and background is red；
E. lines post-processing：According to edge mean breadth, target image is corroded or expansion process, obtain smoother
Profile.
2. image recognition algorithm according to claim 1 is in the application of paper-cut effect, it is characterised in that：The step（1）
In be, by Gaussian filter filtering and noise reduction, cromogram to be subjected to binary conversion treatment, gray-scale map is converted to artwork noise-removed filtering
Picture.
3. image recognition algorithm according to claim 1 is in the application of paper-cut effect, it is characterised in that：The step（2）
In rim detection be that detection is realized using difference in certain picture characteristics of object and background, the difference includes ash
Degree, color or textural characteristics.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710265951.5A CN107038708A (en) | 2017-04-21 | 2017-04-21 | Application of the image recognition algorithm in paper-cut effect |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710265951.5A CN107038708A (en) | 2017-04-21 | 2017-04-21 | Application of the image recognition algorithm in paper-cut effect |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107038708A true CN107038708A (en) | 2017-08-11 |
Family
ID=59536156
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710265951.5A Pending CN107038708A (en) | 2017-04-21 | 2017-04-21 | Application of the image recognition algorithm in paper-cut effect |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107038708A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110032970A (en) * | 2019-04-11 | 2019-07-19 | 深圳市华付信息技术有限公司 | Biopsy method, device, computer equipment and the storage medium of high-accuracy |
| CN110210347A (en) * | 2019-05-21 | 2019-09-06 | 赵森 | A kind of colored jacket layer paper-cut Intelligentized design method based on deep learning |
| CN114612355A (en) * | 2022-03-02 | 2022-06-10 | 影石创新科技股份有限公司 | Oil painting stylization method, device, equipment and medium for image |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101588438A (en) * | 2009-06-10 | 2009-11-25 | 拓维信息系统股份有限公司 | Mobile phone animation creation method for converting digitalized coloured picture into pictures of different styles |
| US20120169759A1 (en) * | 2011-01-04 | 2012-07-05 | Samsung Electronics Co., Ltd. | Method and apparatus for creating a live artistic sketch of an image |
| CN102609964A (en) * | 2012-01-17 | 2012-07-25 | 湖北莲花山计算机视觉和信息科学研究院 | Portrait paper-cut generation method |
| CN103793550A (en) * | 2013-11-14 | 2014-05-14 | 曾凡忠 | Method for quickly making figure paper-cut patterns |
| CN105096358A (en) * | 2015-08-05 | 2015-11-25 | 云南大学 | Line enhanced simulation method for pyrography artistic effect |
| CN105574814A (en) * | 2016-01-06 | 2016-05-11 | 华南理工大学 | Portrait paper-cut special effect generation method |
-
2017
- 2017-04-21 CN CN201710265951.5A patent/CN107038708A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101588438A (en) * | 2009-06-10 | 2009-11-25 | 拓维信息系统股份有限公司 | Mobile phone animation creation method for converting digitalized coloured picture into pictures of different styles |
| US20120169759A1 (en) * | 2011-01-04 | 2012-07-05 | Samsung Electronics Co., Ltd. | Method and apparatus for creating a live artistic sketch of an image |
| CN102609964A (en) * | 2012-01-17 | 2012-07-25 | 湖北莲花山计算机视觉和信息科学研究院 | Portrait paper-cut generation method |
| CN103793550A (en) * | 2013-11-14 | 2014-05-14 | 曾凡忠 | Method for quickly making figure paper-cut patterns |
| CN105096358A (en) * | 2015-08-05 | 2015-11-25 | 云南大学 | Line enhanced simulation method for pyrography artistic effect |
| CN105574814A (en) * | 2016-01-06 | 2016-05-11 | 华南理工大学 | Portrait paper-cut special effect generation method |
Non-Patent Citations (4)
| Title |
|---|
| DANIJELA MARKOVIC 等: "Parameterized Sketches from Stereo Images", 《SPIE-IMAGE & VIDEO COMMUNICATIONS & PROCESSING》 * |
| 朱敬意: "围楼古建筑的三维建模及风格渲染研究", 《中国优秀硕士学位论文全文数据库-信息科技辑》 * |
| 王会芹: "基于图像空间的素描效果生成技术", 《计算机应用》 * |
| 税琳琳 等: "基于图像边界识别的剪纸系统设计", 《中国计算机图形学进展2008-第七届中国计算机图形学大会论文集》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110032970A (en) * | 2019-04-11 | 2019-07-19 | 深圳市华付信息技术有限公司 | Biopsy method, device, computer equipment and the storage medium of high-accuracy |
| CN110210347A (en) * | 2019-05-21 | 2019-09-06 | 赵森 | A kind of colored jacket layer paper-cut Intelligentized design method based on deep learning |
| CN110210347B (en) * | 2019-05-21 | 2021-03-23 | 赵森 | Intelligent color jacket paper-cut design method based on deep learning |
| CN114612355A (en) * | 2022-03-02 | 2022-06-10 | 影石创新科技股份有限公司 | Oil painting stylization method, device, equipment and medium for image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108898610A (en) | A kind of object contour extraction method based on mask-RCNN | |
| CN104751142B (en) | A kind of natural scene Method for text detection based on stroke feature | |
| Zhang et al. | Lighting and pose robust face sketch synthesis | |
| CN109919960B (en) | Image continuous edge detection method based on multi-scale Gabor filter | |
| CN103198319B (en) | For the blurred picture Angular Point Extracting Method under the wellbore environment of mine | |
| CN106446862A (en) | Face detection method and system | |
| CN104794693B (en) | A kind of portrait optimization method of face key area automatic detection masking-out | |
| CN106127193B (en) | A kind of facial image recognition method | |
| CN103218605A (en) | Quick eye locating method based on integral projection and edge detection | |
| CN106446925A (en) | Dolphin identity recognition method based on image processing | |
| CN109426814A (en) | A kind of positioning of the specific plate of invoice picture, recognition methods, system, equipment | |
| CN102024156A (en) | Method for positioning lip region in color face image | |
| CN107945244A (en) | A kind of simple picture generation method based on human face photo | |
| Hagara et al. | About Edge Detection in Digital Images. | |
| Gupta et al. | FPGA based real time human hand gesture recognition system | |
| CN107038708A (en) | Application of the image recognition algorithm in paper-cut effect | |
| CN104331683A (en) | Facial expression recognition method with noise robust | |
| CN102930538A (en) | Adaptive image segmentation method for combining Gaussian potential histogram and spatial histogram | |
| CN106599891A (en) | Remote sensing image region-of-interest rapid extraction method based on scale phase spectrum saliency | |
| Mu et al. | A novel Shi-Tomasi corner detection algorithm based on progressive probabilistic hough transform | |
| CN107784626B (en) | Three-dimensional digital concave-convex relief generating method based on single image | |
| Cheng et al. | Power pole detection based on graph cut | |
| Lin et al. | Machine recognition for broad-leaved trees based on synthetic features of leaves using probabilistic neural network | |
| Sreedevi et al. | Ngfica based digitization of historic inscription images | |
| Rajithkumar et al. | Template matching method for recognition of stone inscripted Kannada characters of different time frames based on correlation 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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170811 |