KR20030057726A - Automatic inbetween generation method using contour matching - Google Patents

Abstract

PURPOSE: A method of automatically generating intermediate motion pictures of contour matching based animation is provided to create natural intermediate motion pictures through a simple method using nonlinear moving lines based on contour matching points. CONSTITUTION: Fractals are extracted from two animation motion pictures. The extracted fractals are combined with each other to generate layers(301). Characteristic points on the contour of each layer are extracted and matched with all points on the contour(302). Linear/nonlinear moving lines are applied to all matching points to generate a contour motion picture(303). Color information of inner points constructing the layer is restored, to generate an intermediate layer(304). All of the generated layers are integrated to generate an intermediate motion picture(305).

Description

Automatic inbetween generation method using contour matching}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for generating an intermediate moving picture of an animation, and more particularly, to an automatic method for automatically generating an intermediate moving picture of an outline matching-based animation for constructing an efficient work environment by automating a moving picture creation process and achieving time and economical savings.

With the development of computers, many attempts have been made to apply classical cell animation techniques to computer-based animation. In particular, 3D animation is being actively researched along with games, and this 3D animation is mainly due to motion capture technology.

Such 3D animation is a database-based image processing technique in which realistic motion data is used to create realistic motions, but motion information must be acquired for new motions.

In addition, 3D animation is mainly represented by skeleton kinematics and inverse kinematics.

On the other hand, unlike 3D animation, 2D animation does not have a large amount of three-dimensional information, and the degree of realistic and three-dimensional description is less than that of 3D animation. Therefore, the use of 3D animation techniques such as kinematics and inverse kinematics for 2D animation is inefficient and inefficient, and much of the information needed to use 3D animation techniques is not considered at all when writing 2D animation. Rather, it comes with overhead.

Due to this problem, 2D animation uses a linear keyframe technique.

Hereinafter, a method for generating an intermediate moving picture of an animation according to the prior art will be described with reference to the accompanying drawings.

First, the method of generating an intermediate fairy tale according to the prior art extracts a skeleton and a joint of a moving object in the first fairy tale and the final fairy tale, and then obtains and locates a center of movement or rotation in the extracted skeleton and joint, and selects a copper wire of each joint. Intermediate assimilation along the movement of each joint is generated.

However, in the conventional method of generating an intermediate fairy tale, as shown in FIG. 1, the shape of an object is distorted in the intermediate fairy tale according to the geometric position of the joint without considering the relation of each joint, and also the intermediate fairy tale using the skeleton. After generating the skeleton of, the process of restoring the plane is very difficult, and there is a problem that there is a discontinuity point due to the quantization error between the points constituting the plane.

Meanwhile, FIG. 2 shows an intermediate animation generated by the kinematic technique and the linear keyframe technique according to the prior art, and it can be seen that a distorted result having a very short form is generated in the intermediate animation of an object such as a rotating arm. Can be.

That is, the conventional intermediate animation generation method can obtain some natural fairy tale using the linear keyframe technique, but it is impossible to obtain a natural fairy tale for the layer of mechanical movement such as the arm and leg of the object. After reconstructing, there is a technique of filling a color inside a character as his normal vector, but it is impossible to automatically generate skeleton data in 2D animation, and there is a problem that it is not suitable for restoring 2D animation because there is too much information.

The present invention has been made to solve the problems of the prior art as described above, and the method of automatically generating an intermediate animation of an edge registration-based animation suitable for generating a natural intermediate animation in a concise manner using a nonlinear copper line based on the boundary registration point. The purpose is to provide.

In order to achieve the above object, the method of automatically generating an intermediate moving picture of an outline matching-based animation of the present invention includes extracting a layer using two originals in an animation production process, extracting feature points on an outline of the layer, Matching points on all outlines, applying linear / nonlinear copper lines to all matching points to create outline assimilation, restoring the interior color to create an intermediate layer, and merging all created layers To generate an intermediate moving picture.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a diagram showing a problem caused by the prior keyframe method according to the prior art.

2 is a view showing the results obtained by the linear keyframe method according to the prior art

3 is a schematic flowchart illustrating a method for automatically generating an intermediate moving picture of an outline matching-based animation according to the present invention.

Figure 4 is a flow chart of the results of various types of image appearing in each step of the present invention

5 shows a general example of a fractal, which is the basic processing element of the present invention;

6 is a flowchart illustrating a method of generating an intermediate moving picture using a nonlinear copper line based on an outline according to the present invention.

7 shows an example of a registration process for an outline that is a key processing element of the present invention.

8 is a view showing a processing example of a nonlinear copper wire processed according to the present invention.

9 illustrates one layer result generated in accordance with the present invention.

10 shows an example of moving pictures generated in accordance with the present invention.

Hereinafter, a method for automatically generating an intermediate moving picture of an outline matching-based animation of the present invention will be described with reference to the accompanying drawings.

First, the method for automatically generating an intermediate moving picture of an outline matching-based animation of the present invention automatically generates an intermediate moving picture simply without using any additional information other than information to be considered when creating an original 2D animation. That is, various additional information used in the 3D animation processing technique such as attribute information, 3D type information, motion and motion characteristic information of the object constituting the animation image is not used at all. This can be said that the image itself of 2D animation is consistent with the basic purpose of conveying simple and characteristic meaning.

In addition, several intermediate animations can be automatically generated using only two essential animated images. In other words, if there is an image in which two moving images do not appear, the moving object cannot be restored without additional information in the intermediate moving picture, so the object disappears and reappears, or three-dimensionally within the image. In the case of rotation, an intermediate moving picture is automatically generated using a moving picture where a new form of the object begins to appear.

In addition, the method of automatically generating the intermediate moving picture of the animation based on the matching line according to the present invention can be easily used by a general person without basic knowledge of image processing by lowering the level of information input by the user as much as possible.

3 is a step-by-step flowchart illustrating a method for automatically generating an intermediate moving picture of an outline matching-based animation according to the present invention.

As shown in the figure, after extracting a fractal from two input animation fairy tale (301), a layer is generated by combining the fractals extracted from each animation fairy tale (302). In this case, the number of layers of each animation moving picture may be different from each other, and there may exist only a layer appearing on either side according to the image. For reference, the fractal may mean an object that forms one region, that is, one unit region, and a plurality of fractals may form one object. In addition, a layer is a unit of motion composed of a plurality of fractals, and thus a layer does not define a single object having meaning, but rather an object having the same movement, so that one layer is clearly distinguished from other layers. It has a hierarchical relationship with the movement characteristics.

Subsequently, in step 302, if a layer for each animation is extracted, a layer outline image is extracted for each layer, and matching and interpolation are performed for all points constituting the outline (303). That is, a point having an outline feature is extracted from the points constituting the outline and matched first, and the remaining points are interpolated in consideration of the relationship with the matched feature point. For reference, outline matching refers to a process of finding a corresponding identical point in the outline of the final fairy tale to all points constituting the outline of the first fairy tale. For reference, the layer outline means an outline that describes the appearance of one layer.

Subsequently, when the outline image of the intermediate fairy tale is generated, color information is restored for all points in the outline in consideration of the geometric positional relationship with the feature points of the outline for all the points constituting the layers in the two animated fairy tales (304). ).

Finally, when all the layers of the intermediate moving picture are generated, one intermediate moving picture is generated by integrating the layer images in consideration of the hierarchical relationship between the layers (305).

This will be described in more detail.

As shown in FIG. 4, two input animation fairy tales composed of the first and last fairy tales are processed first (401, 402), and a plurality of layer images are generated for each input fairy tale through a layer extraction process (403, 404). .

Then, after extracting the outline image of the layer and matching all points on the outline (405, 406), if the number of intermediate moving images to be generated is determined, the layer outline image of the intermediate moving image is considered in consideration of the space-time positional relationship with the two input images. (407, 408, 409).

Subsequently, in consideration of the positional relationship between the two input images and the positional relationship in the layer, color information is restored for the points inside the layer to generate layer images of each intermediate moving image (410, 411, and 412).

Finally, the intermediate moving image is generated by integrating the generated layer images of the respective intermediate moving images according to the hierarchical relationship between the layers (413, 414, 415).

On the other hand, Figure 5 shows a fractal (Fractal) according to the method for automatically generating the intermediate animation of the contour matching-based animation of the present invention, the body 502, the right arm 503, the upper part 504, The left and right boots 505 and 506 become the fractal.

In addition, the face, the head, the hand, the thigh, and the like can all be detected as a fractal, and a combination of a plurality of fractals having a meaning as a moving object can be layered by combining the left thigh with the left boots. Configure

FIG. 6 is a flowchart for explaining in detail a method for automatically generating an intermediate moving picture of an outline matching-based animation according to the present invention.

As shown in the figure, an image of a layer and an outline image of the layer are extracted from the first and last moving images of the input animation (S601).

Thereafter, characteristic points are extracted from the outline images of all the layers (S602), and the characteristic points on the outline of the layer of the last moving picture input image corresponding to the layer of the input first moving picture are matched (S603). In this case, the characteristic point is a point on the outline indicating the outer layer shape and determines a characteristic part of the layer shape, and the matching means finding the same point among the points between two images.

Subsequently, points on the outline between the characteristic points are matched using the matched characteristic points in consideration of a geometric relationship with the characteristic points (S604), and then an intermediate moving picture position to be generated is selected (S605). .

Then, after selecting the corresponding layer of the first moving image input image and the final moving image input image (S606), after determining the copper wire representing the movement of the selected layer (S607), the parameter value for determining the various properties of the selected copper wire is appropriately By selecting (S608), the movement is best represented.

Subsequently, for each point on the matched outline, a point corresponding to the intermediate moving position on the moving line is generated to generate a layer outline image of the intermediate moving image (S609), and then, the layer is configured in consideration of the geometric relationship with the characteristic point. The color information of all internal points is restored (S610).

Subsequently, it is determined whether the generated layered moving picture is appropriate (S611). If it is determined that appropriate, all the layer images are generated through the above process and then all the layer images are generated (S612). If the image is generated, an intermediate moving image is generated by integrating all the generated layer images (S613).

Subsequently, the above process is repeated for all intermediate moving pictures (S614) to complete the animation.

In addition, the restoration of an object existing on one of the first fairy tale and the last fairy tale of the input animation but not on the other is a missing part of the object by matching a point to a plane, a point to a line, and a line to a plane. Automatically generates fairy tales.

On the other hand, Figure 7 is an example of the outline matching process, the core technology of the present invention. When one layer image for two input images is the same as 701 and 702, the outline image is extracted with a line indicating the appearance of each layer.

Among the points constituting the outline, the characteristic points (points 1, 2, and 3) are extracted as points that determine the characteristic shape among the shapes of the outline, and each feature point is joined.

8 is an example showing a representative copper wire proposed in the present invention.

In a typical animation, most of the movement of an object can be represented by a sinusoid and a circle, and these two movements are well depicted by using a single variable (ratio of the length between two registration points) that determines the curve of the movement. .

Meanwhile, FIG. 9 is an example of a layer of intermediate animation generated using the method of automatically generating the intermediate animation of the contour matching-based animation and the sinusoidal copper line, which is different from the result of the conventional technique shown in FIG. It can be seen that the movement is well represented.

FIG. 10 is an example of the final result of an intermediate fairy tale generated in accordance with the present invention. In this example, all 24 intermediate fairy tale are generated, but only eight of them are displayed. You can see that both form and color are naturally well formed.

As described above, the method for automatically generating an intermediate moving picture of an outline matching-based animation of the present invention has the following effects.

First, it provides a function to automatically generate intermediate fairy tales in the process of animation production, and it is possible to obtain a natural level of animation created by hand.

Second, the present invention can reduce the manpower investment cost for animation production. That is, only the core fairy tale is manually created and the remaining intermediate fairy tales are automatically generated by the method proposed in the present invention, which can drastically reduce manpower input costs.

Third, the time required for the completion of animation production can be greatly reduced, and as a result, economic and time input cost reduction can be expected for 2D animation work.

Fourth, it is expected that the development of programs for various uses using the present invention will be possible. That is, educational effects on animations can be expected through development of animation writing programs for students and animation writing programs for the general public, and can be usefully used for homepage production.

Claims (6)

Extracting layers using two originals during the animation process, Extracting feature points on the outline of the layer and matching the feature points with points on all outlines, Generating outline assimilation by applying linear / nonlinear copper to all registration points, Restoring the interior color to create an intermediate layer, Automatically generating the intermediate animation of the contour registration based animation, characterized in that it comprises the step of generating the intermediate animation by integrating all the generated layers. The method of claim 1, wherein the extracting of the layer comprises: Extracting fractals from two input animation fairy tales, And generating a layer by combining the fractals extracted from each of the animated fairy tales. The method of claim 1, wherein the matching of the feature point with the points on all the outlines includes: After extracting a few feature points for the outline of the layer, and matching all the other points on the outline in consideration of the geometric relationship with the few feature points, automatic matching method for generating an intermediate animation of the outline registration-based animation. The method of claim 1, wherein the reconstruction of the internal color information automatically restores the color information for all internal points constituting the layer in consideration of the geometric relationship with the characteristic points. How to create. The method of claim 1, wherein the integration of the layers is performed in consideration of a hierarchical relationship between layers. The contour matching-based animation of claim 1, wherein the reconstruction of an object in which one of the two original paintings is present but the other is not exists matches a point by a plane, a point by a line, and a line by a plane. Automatic generation of intermediate fairy tales.