CN107481184B - Low polygon style diagram generation interactive system - Google Patents

Low polygon style diagram generation interactive system Download PDF

Info

Publication number
CN107481184B
CN107481184B CN201710680738.0A CN201710680738A CN107481184B CN 107481184 B CN107481184 B CN 107481184B CN 201710680738 A CN201710680738 A CN 201710680738A CN 107481184 B CN107481184 B CN 107481184B
Authority
CN
China
Prior art keywords
vertex
vertexes
module
grid
sampling
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201710680738.0A
Other languages
Chinese (zh)
Other versions
CN107481184A (en
Inventor
陈雪锦
马祎婷
白宇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Science and Technology of China USTC
Original Assignee
University of Science and Technology of China USTC
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by University of Science and Technology of China USTC filed Critical University of Science and Technology of China USTC
Priority to CN201710680738.0A priority Critical patent/CN107481184B/en
Publication of CN107481184A publication Critical patent/CN107481184A/en
Application granted granted Critical
Publication of CN107481184B publication Critical patent/CN107481184B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/04—Context-preserving transformations, e.g. by using an importance map
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computer Graphics (AREA)
  • Geometry (AREA)
  • Software Systems (AREA)
  • Processing Or Creating Images (AREA)
  • Image Generation (AREA)

Abstract

The invention discloses a low polygon style diagram generation interactive system, which comprises: the vertex sampling and grid generating module is used for carrying out vertex sampling on the basis of keeping the content structure of the input image, sequencing the vertexes according to the importance of the vertexes by using a self-adaptive thinning algorithm, selecting K top-ranked vertexes from the sequenced vertexes according to the vertex number K fed back by the interaction module, and generating a grid by using Delaunay triangulation; the mesh rendering module is used for coloring the triangular meshes to obtain a low-polygon style graph, wherein the light-dark contrast between adjacent triangles is improved by setting a minimum brightness difference, and related minimum brightness parameters are fed back by the interaction module; and the interaction module is used for realizing the interaction between the system and the user and receiving the vertex number K and the minimum brightness parameter input by the user through a corresponding tool. The system has low computation amount, and can easily and rapidly generate a low polygon style sheet.

Description

Low polygon style diagram generation interactive system
Technical Field
The invention relates to the technical field of image processing, in particular to a low polygon style diagram generation interactive system.
Background
In recent years, low polygon stylized pictures have become increasingly popular in non-photorealistic rendering of photographs because of their 3D abstraction. At present, most methods for generating the low-polygon style diagram need a user to have certain knowledge about 3D modeling and a large amount of workload, and meanwhile, the generated low-polygon style diagram is not ideal in effect. For example, using the software trigriff generation directly, more details are lost. As another example, the disclosed algorithm is guided by a characteristic flow field, and the resulting results retain excessive texture on details such as hair, hat decorations, etc., contrary to the original purpose of low polygonization.
Disclosure of Invention
The invention aims to provide a low-polygon style diagram generation interactive system, which reduces the calculation amount and can easily and quickly generate a low-polygon style diagram.
The purpose of the invention is realized by the following technical scheme:
a low polygon style diagram generation interaction system, comprising: the system comprises a vertex sampling and grid generating module, a grid rendering module and an interaction module; wherein:
the vertex sampling and grid generating module is used for performing vertex sampling on the basis of keeping the content structure of the input image, sequencing the vertexes according to the importance of the vertexes by using a self-adaptive thinning algorithm, selecting K top-ranked vertexes from the sequenced vertexes according to the vertex number K fed back by the interaction module, and generating a grid by using Delaunay triangulation;
the mesh rendering module is used for coloring the triangular meshes to obtain a low-polygon style graph, wherein the light-dark contrast between adjacent triangles is improved by setting a minimum brightness difference, and related minimum brightness parameters are fed back by the interaction module;
and the interaction module is used for realizing the interaction between the system and the user and receiving the vertex number K and the minimum brightness parameter input by the user through a corresponding tool.
The interaction module comprises two sliding tool bars; the two sliding tool bars are respectively used for realizing user input and adjusting the vertex number K and the minimum brightness parameter.
The interaction module further comprises a brush tool;
and the brush tool is used for realizing the labeling of the important region, and when the interactive module acquires the important region labeled by the user through the brush tool, the important region is fed back to the vertex sampling and grid generating module, so that the vertex sampling and grid generating module reserves all vertexes L in the important region, and then K-L vertexes before ranking are selected from the sequenced vertexes.
The vertex sampling on the basis of keeping the content structure of the input image and sequencing the vertexes according to the importance of the vertexes by using an adaptive thinning algorithm comprises the following steps:
an input image is divided into superpixels by adopting an over-division method, the intersection points of adjacent superpixels are selected as vertexes, and a vertex candidate set p is formedsubset＝p1,...,pM；
The vertex is the difference from the input image after removing the point in the triangular mesh according to the importance of the vertex; when a vertex is removed, the color of the obtained triangularization result is also changed, and after the vertex is removed, the larger the color difference is, the more important the vertex is; the sorting process is as follows: using Delaunay triangulation to generate a color mesh from the vertex candidate set; recalculating approximation error of each vertex, wherein the smaller the approximation error is, the lower the importance is, and selecting the minimum point p*Add it to the queue; once point p*Removing from the vertex candidate set by point p*Local updating to obtain a new triangulated mesh; point p*The approximation error of neighboring points will be recalculated for use in the next round of point removal; and finishing the sequencing of the M vertexes by continuously repeating the process.
The coloring of the triangular meshes to obtain the low polygon style diagram comprises the following steps:
for a triangular mesh
Figure BDA0001375368510000021
The average value c of the middle part of the corresponding triangle color histogram is selectediFor each set of adjacent triangles TiAnd TjA difference d in luminance between themi,jComprises the following steps:
di,j＝max(|Li-Lj|,Lmin)；
wherein L isi、LjAre respectively adjacent triangles Ti、TjL vector in LAB color space, LminMinimum brightness fed back for the interaction module;
diagonal grid
Figure BDA0001375368510000023
All triangles inL 'is calculated by solving a system of linear equations'i：
Figure BDA0001375368510000022
According to the technical scheme provided by the invention, on one hand, after the sampled vertexes are sequenced, when a user needs to change the number K of the vertexes or mark an important area, the vertexes can be directly called from the sequenced vertexes without recalculation, so that the calculation amount is reduced. Meanwhile, the brightness-dark contrast between adjacent triangles can be improved according to the minimum brightness parameter input by the user, so that the rendered result has a standard deviation larger than that of the rendered result in a color domain in other modes. On the other hand, the two parts of operations can be participated by the user, and the user can easily and quickly generate the low-polygon style sheet through the system.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on the drawings without creative efforts.
FIG. 1 is a schematic diagram of a low polygon style diagram generation interactive system according to an embodiment of the present invention;
FIG. 2 is a flow chart of generating a low polygon style sheet according to an embodiment of the present invention;
FIG. 3 is a diagram illustrating the result of over-segmentation and the initial point set according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating partial triangularization updates after two vertices are removed according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of the adjustment of the number of vertices and the triangulation thereof using a slide bar and a brush tool according to an embodiment of the present invention
FIG. 6 is a rendering result of different color contrasts under the same triangular mesh provided by the embodiment of the present invention;
FIG. 7 is a comparison graph of low polygon styles generated by various schemes provided by embodiments of the present invention;
FIG. 8 is a block diagram illustrating various types of low polygon style sheets generated by the system according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention are clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment of the invention provides a low polygon style graph generation interaction system, which comprises: the system comprises a vertex sampling and grid generating module, a grid rendering module and an interaction module; wherein:
the vertex sampling and grid generating module is used for performing vertex sampling on the basis of keeping the content structure of the input image, sequencing the vertexes according to the importance of the vertexes by using a self-adaptive thinning algorithm, selecting K top-ranked vertexes from the sequenced vertexes according to the vertex number K fed back by the interaction module, and generating a grid by using Delaunay triangulation;
the mesh rendering module is used for coloring the triangular meshes to obtain a low-polygon style graph, wherein the light-dark contrast between adjacent triangles is improved by setting a minimum brightness difference, and related minimum brightness parameters are fed back by the interaction module;
and the interaction module is used for realizing the interaction between the system and the user and receiving the vertex number K and the minimum brightness parameter input by the user through a corresponding tool.
In the embodiment of the invention, the interaction module comprises two sliding tool bars; the two sliding tool bars are respectively used for realizing user input and adjusting the vertex number K and the minimum brightness parameter.
In addition, the interaction module further comprises a brush tool; and the brush tool is used for realizing the labeling of the important region, and when the interactive module acquires the important region labeled by the user through the brush tool, the important region is fed back to the vertex sampling and grid generating module, so that the vertex sampling and grid generating module reserves all vertexes L in the important region, and then K-L vertexes before ranking are selected from the sequenced vertexes.
In the embodiment of the present invention, the sampling vertices on the basis of maintaining the content structure of the input image, and sorting the vertices according to their importance by using an adaptive thinning algorithm includes:
an input image is divided into superpixels by adopting an over-division method, the intersection points of adjacent superpixels are selected as vertexes, and a vertex candidate set p is formedsubset＝p1,...,pM；
The importance of the vertex is the distance from the input image after removing this point in the triangular mesh; when a vertex is removed, the color of the obtained triangularization result is also changed, and after the vertex is removed, the larger the color difference is, the more important the vertex is; the sorting process is as follows: using Delaunay triangulation to generate a color mesh from the vertex candidate set; recalculating approximation error of each vertex, wherein the smaller the approximation error is, the lower the importance is, and selecting the minimum point p*Add it to the queue; once point p*Removing from the vertex candidate set by point p*Local updating to obtain a new triangulated mesh; point p*The approximation error of neighboring points will be recalculated for use in the next round of point removal; and finishing the sequencing of the M vertexes by continuously repeating the process.
In the embodiment of the present invention, the obtaining of the low polygon style diagram by coloring the triangular meshes includes:
for a triangular mesh
Figure BDA0001375368510000041
The average value c of the middle part of the corresponding triangle color histogram is selectediFor each set of adjacent triangles TiAnd TjA difference d in luminance between themi,jComprises the following steps:
di,j＝max(|Li-Lj|,Lmin)；
wherein L isi、LjAre respectively adjacent triangles Ti、TjL vector in LAB color space, LminMinimum brightness fed back for the interaction module;
for triangular mesh
Figure BDA0001375368510000042
All triangles of (1) whose enhanced luminance L 'is calculated by solving a system of linear equations'i：
Figure BDA0001375368510000043
For the sake of understanding, the following describes the above system in detail with reference to the accompanying drawings.
As shown in fig. 2, for a flow chart for generating a low polygon style sheet using the system of fig. 1,
the method comprises the following steps:
step a, inputting an image.
Suppose that the input image I contains N pixels p1,...pNThe object of the invention is to generate a low polygon consisting of triangular meshes
Figure BDA0001375368510000051
And each triangle has a particular color. A good low polygon requires a structure similar to the input image I at an abstract level and has a certain artistic quality.
And b, sampling the top point.
In order to allow the user to select the number of vertices easily and efficiently, we do not need to repeat the sampling operation since each sampling must have many important points to be sampled in order to maintain the content structure of the picture. To reduce computational complexity, we first extract important, valid information-containing points from the input picture and sort them by their importance in ensuring the structural integrity of the picture content. Once an exact number of vertices K is determined, we simply select the K top ranked and use Delaunay triangulation to generate the mesh.
The step can be divided into sub-steps of determining candidate vertexes and sequencing vertexes; the method comprises the following specific steps:
a1, determining candidate vertexes
In order to effectively retain structural information of an input image and enable a sampling point to express the outline of the image, firstly, an over-segmentation method is adopted to segment the image into superpixels, then, intersections of adjacent superpixels are selected as initial vertex candidate vertexes, and therefore a candidate vertex set P is obtainedM＝p1,...,pMWherein, K is less than M and less than N.
There are many over-segmentation algorithms in the relevant literature, and we choose the SLIC algorithm in view of the edge-preserving capability and its algorithmic efficiency. First, to obtain as much structural information as possible, we have chosen a significantly larger value M of 3000 superpixels. Fig. 3 illustrates the result of over-segmentation (i.e., fig. 3a) and the initial set of points (i.e., fig. 3b), it can be seen that the selected points are located at the edge of the object and can approximately generate an evenly distributed grid.
a2, vertex order
After the segmentation operation, the number of sampling points is too large for generating the low-polygon picture. Given a certain number of vertices, K, we need to select K points from the initial M points to construct the triangular mesh. However, if the user wants to choose the value of K within a large range at will, it is not necessary to run the vertex sampling step each time. Therefore, the invention provides a ranking strategy adopting an adaptive thinning algorithm, and points with lower importance can be removed circularly according to some judgment criteria depending on data. The M points only need to be ranked once, and according to the ranking, a user selects a K value, and the system can directly generate a corresponding triangular mesh. This is different from previous methods in that when the user wishes to change the value of K, the entire process does not need to be re-run, greatly improving computational efficiency.
In the embodiment of the present invention, the criterion of the importance of a vertex is the difference from the original image after removing this point in the triangular mesh M. When a dot is removed, the color of the resulting triangulated result also changes. After removing a point, the larger the color difference, the more important this point is. For each triangle T in the triangular mesh MiTo avoid aliasing, the average value of the middle part of the color histogram in the corresponding position of the input image is selected.
If a point set P of k pointskHas been selected, and according to these points, the Delaunay triangulation is carried out on the picture to obtain a subdivision grid D (P)k). For D (P)k) The average color value of the original image I in the triangle is calculated for each triangle, and the average color value is given to the triangle, thereby generating a color mesh image
Figure BDA0001375368510000061
To approximate the input image I. In summary, for the point set PkObtaining a color grid image
Figure BDA0001375368510000062
Its approximation error with respect to the input image I
Figure BDA00013753685100000612
Is the color difference of all pixels x in the image:
Figure BDA0001375368510000063
wherein,
Figure BDA0001375368510000064
is the color value of the input image at pixel point x, and Mx(Pk) It is a color grid image
Figure BDA0001375368510000065
The color value at pixel point x. After point P is removed, a new triangularization D (P) is constructedk-p) andregenerating the colors of the triangles to obtain another color mesh
Figure BDA0001375368510000066
Can calculate
Figure BDA0001375368510000067
Has an approximation error of
Figure BDA0001375368510000068
For slave colour grids according to Delaunay criterion
Figure BDA0001375368510000069
To the colour grid
Figure BDA00013753685100000610
Is local, which means that only the local area where the point p exists needs to be re-triangulated and the local approximation error is calculated as:
Figure BDA00013753685100000611
where N (p) refers to the local area formed by the triangle with the vertex p.
From the approximation error for each point, the point with the smallest approximation error can be removed in each cycle.
The above process is illustrated in FIG. 4, which shows a schematic diagram of partial triangularization update after removing two vertices. Both the top and bottom panels of fig. 4a are color meshes obtained by delaunay triangularization and averaging the triangles. FIG. 4b is a schematic diagram of the top and bottom diagrams showing the triangularization update only in a local area after a vertex is removed; FIG. 4c is a top and bottom graph showing color differences after triangularization update; fig. 4d shows an input image. From the top four images of fig. 4, in the smoother area, the color change is not obvious after removing one point; from the lower four images in fig. 4, when the removed point is in a place with rich structural information, the removed point has a larger difference compared with the original image; in other words, the removed vertices in the top image in FIG. 4b are less important than the removed vertices in the bottom image.
Based on the above principle, for obtaining candidate vertex sets, the color mesh is generated by using Delaunay triangulation
Figure BDA0001375368510000071
Recalculating approximation error of each vertex, wherein the smaller the approximation error is, the lower the importance is, and selecting the minimum point p*Add it to the queue; once point p*Removing from the vertex candidate set by point p*Local updating to obtain a new triangulated mesh; point p*The approximation error of neighboring points will be recalculated for use in the next round of point removal; and finishing the sequencing of the M vertexes by continuously repeating the process.
After ranking all the M vertexes, a user can randomly designate a K value through the interaction module, and the system can directly select the vertexes K before ranking to perform subsequent grid generation and rendering work. Illustratively, here K ∈ (100, M ].
Of course, the user can also mark important areas by using a brush tool in the interactive module, for example, in the portrait, the eyes, nose and other five sense organs of the character have more details and need more points to express than the shoulders and the like. Therefore, the user can draw the regions which are considered to be important and need to be highlighted on the original image by using the brush provided by the system, and because the importance of the points is calculated only once by using the algorithm, the system can reserve all initial points in the regions and correspondingly delete other relatively unimportant points, namely, the system can reserve all vertexes L in the important regions and then select K-L top points before ranking from the sorted vertexes.
The number of sampled vertices has a large impact on the effect of the finally generated low-polygon stylist. In the previous method, a vertex number is directly determined according to the size of an original image. However, when pictures with more detail or more complex texture are encountered, the results obtained with this approach may not be necessarily satisfactory. In this case, it is important to change the number of vertices in real time and observe the result. The method for specifying the K value or specifying the K value and marking the important area can be realized through a sliding bar and a brush tool. The sliding bar can allow the user to continuously drag and directly observe the corresponding result.
FIG. 5 is a schematic diagram of the adjustment of the number of vertices and their triangularization using a slider and brush tool. K in fig. 5a is 500; k750 in fig. 5 b; FIG. 5c is the triangularization result of FIG. 5 b; fig. 5d is a schematic diagram of drawing the important region when K is 750; FIG. 5e is the vertex selection result of FIG. 5 d; FIG. 5f shows the triangularization result of FIG. 5 e.
And c, generating a grid.
And carrying out Delaunay triangulation according to the finally selected vertex to obtain a triangular mesh.
And d, grid rendering.
After the triangle mesh is generated, the next step is to color the triangles to obtain a more visually artistic low polygon style sheet. One simple approach is to take the average color of each triangle. However, low polygon stylization is characterized by its visual 3D effect, which is manifested by shadows on the geometric surface. In order to achieve this effect as much as possible, the embodiment of the present invention improves the contrast between light and dark between adjacent triangles by setting the minimum luminance difference.
For a triangular mesh
Figure BDA0001375368510000083
The average value c of the middle part of the corresponding triangle color histogram is selectediFor each set of adjacent triangles TiAnd TjA difference d in luminance between themi,jComprises the following steps:
di,j＝max(|Li-Lj|,Lmin)；
wherein L isi、LjAre respectively adjacent triangles Ti、TjL vector in LAB color space, LminMinimum brightness fed back for the interaction module;
diagonal grid
Figure BDA0001375368510000081
All triangles of (1) whose enhanced luminance L 'is calculated by solving a system of linear equations'i：
Figure BDA0001375368510000082
As shown in fig. 6, the rendering results of different color contrasts under the same triangular mesh are shown. The average value of each triangle color in the two graphs of FIG. 6a is taken; fig. 6b to fig. 6c all use the above solution of the embodiment of the present invention to improve the light and dark contrast of the adjacent triangles.
As shown in fig. 6, the rendering method provided by the embodiment of the present invention has a larger standard deviation in the color gamut. Meanwhile, a sliding bar is provided in the interactive system, and a user can observe different effects under different color contrasts in real time by dragging the sliding bar and select a more inclined result. Thus, there may be more choices for the user.
On the other hand, in order to illustrate the effects of the above-mentioned system provided by the embodiment of the present invention, a comparison is also made with the prior art.
FIG. 7 shows the results of algorithm processing covered in the paper published in 2015 in FIG. 7a and the results of software TRIGRAFF in FIG. 7 b; FIGS. 7 c-7 d show the results of the above system according to the embodiment of the present invention. It can be seen that the two images of fig. 7a retain excessive texture in details such as hair, cap decoration, etc., contrary to the original intention of low polygonization. The results generated by the two images of fig. 7b are not ideal, especially in terms of detail. And fig. 7c to 7d not only can retain more details, but also can perform more triangularization and simplification like artists on some excessively complicated details. It can be obviously seen that the effect of the low polygon style diagram generated by the system provided by the embodiment of the invention is obviously better than the generation results of the other two schemes. Fig. 8 shows more results on different types of pictures, and it can be seen that our system has a good performance on different styles of pictures, such as objects, people, landscapes, cartoons, etc.
In addition, in order to embody the efficiency of the algorithm, a relevant calculation time statistical table is also listed. The program runs in the following environment: intel i7 CPU, 3.4GHZ, 16GB memory. As shown in table 1. For the initialization of M-2000 points, the adaptive refinement process (i.e., vertex sorting process) takes on average about 3.5 seconds, and the rest of the process can be performed in real time.
Examples of the present invention Size of picture Over-partition time Sequencing time Rendering time
Flower (Picture 5) 614*410 0.156 3.009 0.059
Ice cream (Picture 4) 444*507 0.136 3.129 0.044
Lina (FIG. 6) 512*512 0.187 3.363 0.062
Hamburger (fig. 7) 717*537 0.239 3.959 0.086
Through the above description of the embodiments, it is clear to those skilled in the art that the above embodiments can be implemented by software, and can also be implemented by software plus a necessary general hardware platform. With this understanding, the technical solutions of the embodiments can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a usb disk, a removable hard disk, etc.), and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods according to the embodiments of the present invention.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims (4)

1. A low polygon style drawing generation interactive system, comprising: the system comprises a vertex sampling and grid generating module, a grid rendering module and an interaction module; wherein:
the vertex sampling and grid generating module is used for performing vertex sampling on the basis of keeping the content structure of the input image, sequencing the vertexes according to the importance of the vertexes by using a self-adaptive thinning algorithm, selecting K top-ranked vertexes from the sequenced vertexes according to the vertex number K fed back by the interaction module, and generating a grid by using Delaunay triangulation;
the mesh rendering module is used for coloring the triangular meshes to obtain a low-polygon style graph, wherein the light-dark contrast between adjacent triangles is improved by setting a minimum brightness difference, and related minimum brightness parameters are fed back by the interaction module;
the interaction module is used for realizing the interaction between the system and the user and receiving the vertex number K and the minimum brightness parameter input by the user through a corresponding tool;
wherein, the sampling the vertexes on the basis of keeping the content structure of the input image, and ordering the vertexes according to the importance thereof by using the self-adaptive thinning algorithm comprises the following steps:
an input image is divided into superpixels by adopting an over-division method, the intersection points of adjacent superpixels are selected as vertexes, and a vertex candidate set p is formedsubset＝p1,...,pM；
The vertex is the difference from the input image after removing the point in the triangular mesh according to the importance of the vertex; when a vertex is removed, the color of the obtained triangularization result is also changed, and after the vertex is removed, the larger the color difference is, the more important the vertex is; the sorting process is as follows: using Delaunay triangulation to generate a color mesh from the vertex candidate set; recalculating approximation error of each vertex, wherein the smaller the approximation error is, the lower the importance is, and selecting the minimum point p*Add it to the queue; once point p*Removing from the vertex candidate set by point p*Local updating to obtain a new triangulated mesh; point p*The approximation error of neighboring points will be recalculated for use in the next round of point removal; and finishing the sequencing of the M vertexes by continuously repeating the process.
2. The low polygon style drawing generation interaction system of claim 1, wherein the interaction module comprises two sliding toolbars; the two sliding tool bars are respectively used for realizing user input and adjusting the vertex number K and the minimum brightness parameter.
3. A low polygon style drawing generation interactive system according to claim 1 or 2, wherein the interactive module further comprises a brush tool;
and the brush tool is used for realizing the labeling of the important region, and when the interactive module acquires the important region labeled by the user through the brush tool, the important region is fed back to the vertex sampling and grid generating module, so that the vertex sampling and grid generating module reserves all vertexes L in the important region, and then K-L vertexes before ranking are selected from the sequenced vertexes.
4. The system of claim 1, wherein the rendering the triangle mesh to obtain the low polygon style map comprises:
for a triangular mesh
Figure FDA0002480161020000021
The average value c of the middle part of the corresponding triangle color histogram is selectediFor each set of adjacent triangles TiAnd TjA difference d in luminance between themi,jComprises the following steps:
di,j＝max(|Li-Lj|,Lmin)；
wherein L isi、LjAre respectively adjacent triangles Ti、TjL vector in LAB color space, LminMinimum brightness fed back for the interaction module;
diagonal grid
Figure FDA0002480161020000022
All triangles of (1) whose enhanced luminance L 'is calculated by solving a system of linear equations'i：
Figure FDA0002480161020000023
CN201710680738.0A 2017-08-10 2017-08-10 Low polygon style diagram generation interactive system Active CN107481184B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710680738.0A CN107481184B (en) 2017-08-10 2017-08-10 Low polygon style diagram generation interactive system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710680738.0A CN107481184B (en) 2017-08-10 2017-08-10 Low polygon style diagram generation interactive system

Publications (2)

Publication Number Publication Date
CN107481184A CN107481184A (en) 2017-12-15
CN107481184B true CN107481184B (en) 2020-08-25

Family

ID=60599145

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710680738.0A Active CN107481184B (en) 2017-08-10 2017-08-10 Low polygon style diagram generation interactive system

Country Status (1)

Country Link
CN (1) CN107481184B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11010963B2 (en) * 2018-04-16 2021-05-18 Nvidia Corporation Realism of scenes involving water surfaces during rendering

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103164864A (en) * 2011-12-13 2013-06-19 上海炬力集成电路设计有限公司 Triangularization method of polygon in computer image processing and system using the same
CN105989584A (en) * 2015-01-29 2016-10-05 北京大学 Image stylized reconstruction method and device

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103164864A (en) * 2011-12-13 2013-06-19 上海炬力集成电路设计有限公司 Triangularization method of polygon in computer image processing and system using the same
CN105989584A (en) * 2015-01-29 2016-10-05 北京大学 Image stylized reconstruction method and device

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
AN INTERACTIVE SYSTEM FOR LOW-POLY ILLUSTRATION GENERATION FROM IMAGES USING ADAPTIVE THINNING;Yiting Ma 等;《Proceedings of the IEEE International Conference on Multimedia and Expo (ICME) 2017》;20170714;第1033-1038页 *

Also Published As

Publication number Publication date
CN107481184A (en) 2017-12-15

Similar Documents

Publication Publication Date Title
CN112669447B (en) Model head portrait creation method and device, electronic equipment and storage medium
US20180144509A1 (en) Systems and Methods for Providing Convolutional Neural Network Based Image Synthesis Using Stable and Controllable Parametric Models, a Multiscale Synthesis Framework and Novel Network Architectures
US6987520B2 (en) Image region filling by exemplar-based inpainting
US8917282B2 (en) Separating water from pigment in procedural painting algorithms
Gerstner et al. Pixelated image abstraction with integrated user constraints
US20110050685A1 (en) Image processing apparatus, image processing method, and program
US8406566B1 (en) Methods and apparatus for soft edge masking
Wang et al. Unsupervised texture transfer from images to model collections.
Dong et al. Image retargeting by texture-aware synthesis
WO2019113888A1 (en) Interactive chinese ink painting style stroke generation method and device
Lee et al. Directional texture transfer with edge enhancement
WO2023093291A1 (en) Image processing method and apparatus, computer device, and computer program product
CN108171650B (en) Chinese flower water-ink painting style stroke generation method with stroke optimization function
Liu et al. Generating classic mosaics with graph cuts
Zhang et al. Imageadmixture: Putting together dissimilar objects from groups
CN109448093B (en) Method and device for generating style image
KR101191319B1 (en) Apparatus and method for painterly rendering based on objective motion information
CN107481184B (en) Low polygon style diagram generation interactive system
Kasao et al. Algorithmic Painter: a NPR method to generate various styles of painting
KR100602739B1 (en) Semi-automatic field based image metamorphosis using recursive control-line matching
CN108171766B (en) Image generation method with stroke contour correction function
KR101098830B1 (en) Surface texture mapping apparatus and its method
CN114596213A (en) Image processing method and device
KR101264358B1 (en) Method and System for Automated Photomosaic Image Generation
Chang et al. Artistic painting style transformation using a patch-based sampling method

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