KR100945037B1 - Method and system for creating exaggeration image using rubber-like exaggeration algorithm - Google Patents

Abstract

The present invention discloses an exaggerated image generation method and system that can be used for cartoon animation or character animation. The exaggerated image generation method of the present invention comprises the steps of receiving motion information for a plurality of joints, generating motion information with exaggerated positions through exaggerated position of the joints based on the trajectory of the joints, between adjacent joints. Generating exaggerated motion information for the sub-joints located in and generating an exaggerated image using the exaggerated motion information for the sub-joints. According to the present invention, it is possible to exaggerate the motion free from the limitation of the link length that has been basically observed in the character animation, and to generate an exaggerated image having a flexible and dynamic motion that is bent in a rubber shape.

Motion Exaggeration, Trajectory Exaggeration, Cartoon Animation, Mass-Spring Simulation

Description

Method and system for creating exaggeration image using rubber-like exaggeration algorithm

The present invention relates to a method and a system for generating an exaggerated image, and more particularly, to a method for generating an exaggerated image in a rubber form capable of converting motion capture information into a cartoon style character motion. The invention can be used in the field of cartoon animation, character animation of computer games.

The use of motion capture data has been an important issue in the field of character animation, and the motion graph has been used to produce relatively good results. In particular, physics-based motion editing techniques have been studied by many researchers. Liu and Popovic used dynamic formulas in “In processings of ACM SIGGRAPH '02, pages 408–416, ACM Press, 2002”. We proposed a motion conversion method for physically realistic motion. This document has introduced a physically based representation of realistic motion using nonlinear inverse optimization. In addition, several documents have disclosed a representation method of mixing physical simulation with motion capture.

On the other hand, because cartoon animation is a form of a stylized motion, it is an important issue to style character motion in cartoon animation. This has achieved some good results, but not particularly for cartoons. Changing the goal setting of one character's motion to another involves the problem of mapping one character motion to another character with a different link length. Previously, when dealing with the conceptual approach to changing the link length of the character, there was a case where the skinning method of automatically adding the joint to the original to interpolate the rotation of the joint.

Research into non-photorealistic animation can be classified into visual-dependent approaches, techniques for recycling cartoon animation data, and techniques that follow the principles of traditional animation. Rademacher observes that cartoons can have different representations depending on the field of view they see, and suggests a visual-dependent technique using linear interpolation and that this approach can be applied to animation. Similar approaches can be applied to character animation. Braggler et al. 2 proposed a way to acquire a cartoon animation and reuse it. Braggler et al. 2 attempted to capture squash and stretch effects from cartoon images using affine transformations. However, this method has a certain limitation in exaggeration because it has a rigid link structure.

Many of the methods mentioned above have good results in live or non-realistic animation. However, conventional motion capture techniques cannot directly generate cartoon-style animation. Conventionally, there have been certain constraints on the joint hierarchy of a character, such as maintaining a link at a predetermined length, and this constraint has limited the exaggeration of motion. Motion graphs, a technique for generating realistic motion data for character animation, and physics-based character animination have produced good results with respect to actual motion data. While these methods are great for creating realistic animations, they are not suitable for creating cartoon animations that require different motions depending on the limitations of traditional animations such as joint length and angle constraints. not.

In view of the above limitations of the prior art, the present invention is free from the motion exaggeration, and the flexible and dynamic motion exaggeration in the form of rubber is possible because the present invention is not limited by the link length that has been basically observed in the character animation. An object of the present invention is to provide an exaggerated image generation method and system that can be applied to animation.

An exaggerated image generating method according to the present invention for solving the above technical problem comprises the steps of receiving motion information including position information and direction information for a plurality of joints; Generating exaggerated motion information for the plurality of joints through exaggerated position of the joint based on a trajectory of the joint; Using the exaggerated motion information and the input motion information, generating exaggerated motion information on sub-joints positioned between adjacent joints; And generating an exaggerated image using the exaggerated motion information on the generated joints.

According to another aspect of the present invention, there is provided an exaggerated image generating system comprising: an input unit configured to receive motion information including position information and direction information of a plurality of joints; A motion exaggeration unit configured to generate exaggerated motion information for the plurality of joints through exaggerated position of the joint based on a trajectory of the joint; By using the exaggerated motion information and the motion information input through the input unit, to the limb joint motion exaggeration and the generated sub-joints to generate exaggerated motion information for the sub-joints located between the neighboring joints And an exaggerated image generating unit generating an exaggerated image by using the exaggerated motion information.

The present invention also provides a computer-readable recording medium having recorded thereon a program for executing the above-mentioned exaggerated image generating method on a computer.

According to the present invention, motion exaggeration based on the trajectory of the motion obtained through the mode capture is possible to exaggerate motion free from the limitation of the link length that has been basically observed in the character animation, and through the exaggeration using the joint By modifying the structure of the exaggerated image having a flexible and dynamic motion bending in the form of rubber can be generated.

Hereinafter, a method and system for generating an exaggerated image using a rubber exaggerated algorithm of the present invention and a recording medium on which a program for performing the exaggerated image generating method on a computer will be described in detail with reference to the accompanying drawings and embodiments.

1 is a conceptual diagram for explaining an exaggerated image generation algorithm of the present invention. The exaggerated image generation algorithm according to the present invention is largely divided into two algorithms. One is trajectory-based motion exaggeration, and the other is rubber-joint group smoothing (RJG smoothing) to achieve the bend effect using the sub-articular joints. to be.

First, the trajectory-based motion exaggeration introduced in the present invention is an exaggeration inputting motion capture data having realistic motion, and the trajectory-based motion exaggeration is free from the link length, and thus has various types of motions suitable for cartoon animation. Video production is possible. In the present invention, the trajectory-based motion exaggeration includes motion exaggeration using curve extrapolation and joint hierarchy correction for correcting the joint structure, which will be described later.

The joint group smoothing algorithm involves creating a joint joint layer that splits the link of the original joint structure into small unit joints, or sub-joints. The joints are used here to mimic the bending effect of rubber. The joint structure for generating the joint may be predefined and calculated.

RJG positioning groups the generated joints according to predetermined criteria, smoothes them using multi-dimensional curve interpolation, and then positions the joints according to the joint group smoothing algorithm. For example, the left arm, right arm, left leg, and right leg can be determined for each RJG. Bezier curves can be used for multidimensional curve interpolation.

Mass-spring simulation is a dynamic simulation that identifies each joint and joint as a mass point and assumes that joints and joints adjacent to each other are connected by a spring. This allows you to create more dynamic and fun motions.

Blending is the blending of simulation results and results from RJG positioning, that is, the position information of joints and joints, and post-processing is the new skinning information and direction information of joints whose location is specified by blending. Is generated, and the final motion information is output using the same. In this embodiment, the basic skeletal representation used most widely for character animation is used.

The invention is based on trajectory-based motion exaggeration that does not obey joint length limitations, the introduction of sub-joints more closely to the original joints, and the position of the joints through Bezier curve interpolation and mass-spring simulations. By computing the method of generating an exaggerated image having a smooth movement of the rubber form and dynamic exaggerated motion. Existing approaches only modify the angle of the joint and the like while maintaining the length of the link, but the present invention enables smooth and varied motion exaggeration with a rubbery pose through an approach to modify the structure of the joint itself.

2 is a block diagram illustrating an exaggerated image generating system according to an embodiment of the present invention. The exaggerated image generating system 1 shown in FIG. 2 includes an input unit 10, a motion exaggeration unit 20, a motion correction unit 30, a joint joint generation unit 40, a joint joint exaggeration unit 50, and a mass. A spring simulation unit 60 and an exaggerated image generating unit 70.

The input unit 10 receives motion information about the character image. The motion information includes local coordinate data including position data and orientation information about the plurality of joints.

The motion exaggeration unit 20 generates exaggerated motion information that exaggerates the positions of the plurality of joints by exaggerating the position of the joint based on the trajectory of the joint. In the present embodiment, the motion exaggeration 20 exaggerates motion using curve extrapolation, and a detailed algorithm will be described later.

The motion correction unit 30 generates modified motion information by modifying the exaggerated motion information in consideration of link lengths between neighboring joints in order to prevent image distortion due to the exaggeration.

The sub-joint creation unit 40 defines sub-joints positioned between joints adjacent to each other according to the input motion information and a predetermined criterion.

The joint motion exaggeration unit 50 generates exaggerated motion information on the joints located between the joints adjacent to each other using the exaggerated motion information and the motion information input through the input unit. The joint motion exaggeration unit 50 includes a grouping unit 52, a function estimating unit 54, and a joint joint motion information generation unit 56. This embodiment treats joints and sub-joints in terms of parent joints and child joints between the neighboring joints. The parent joint is a joint located relatively inward of the two joints adjacent to each other, and the child joint is a joint located relatively outward. For example, when the character is a person, the center of the human body may be defined as the inside.

The grouping unit 52 groups joints in which a link between the parent joint having one child joint and the child joint is continuously formed among the plurality of joints. The function estimator 54 estimates a function of connecting joints neighboring each other through curve interpolation for joints included in one group by grouping. The joint motion information generator 56 determines an exaggerated position of the joints using the estimated function and the joint definition criteria used in the joint generator, and generates the exaggerated motion information of the joints.

The mass-spring simulation unit 60 modifies the exaggerated motion information on the sub-joints through mass-spring simulation. In this case, the mass-spring simulation is a simulation by treating the joints and the joints as mass materials having respective masses, and treating the joint forces or the joints and the joints as adjacent to each other.

The exaggerated image generator 70 generates an exaggerated image by using motion information generated from the joint motion information generator 56 and the mass-spring simulation unit 60. The exaggerated image generator 70 includes a mixer 72 and a post processor 74. The mixing unit 72 mixes the motion information modified according to the mass-spring simulation and the exaggerated motion information for the sub-joints generated in the sub-joint motion exaggeration according to a predetermined weight function. The post processor 74 modifies skinning data of the original input image by using the mixed motion information, and generates an exaggerated image according to the modified skinning information. When the direction information is not included in the mixed motion information, the direction information of the exaggerated joints and the sub-joints may be estimated using the direction information of the original joints and the sub-joints that are not exaggerated.

3 is a flowchart illustrating a method of generating an exaggerated image according to an embodiment of the present invention. The exaggerated image generation method illustrated in FIG. 3 includes the following steps performed in time series in the exaggerated image generation system 1.

In step 102, the input unit 102 receives motion information about a plurality of joints. The motion information input in this step is motion information that is not exaggerated. In the present embodiment, motion information is used as a meaning including position information and direction information of joints according to a skeleton representation generally used for character animation.

In step 104, the motion exaggeration unit 104 generates exaggerated motion information through exaggeration of the position of the joint based on the trajectory of the joint. For example, the motion exaggeration 104 generates exaggerated motion information using curve extrapolation. In the conventional motion exaggeration, there is an example in which the joint angle is exaggerated, but in this embodiment, the motion exaggeration is characterized by exaggerating the motion through the trajectory exaggeration of the joint. Existing quaternion-based motion editing does not have a normality problem, but using the above method, the resulting motion with a stretched link cannot be obtained. In this embodiment, a method of exaggerating the trajectory of each joint in the three-dimensional space is introduced, thereby making it possible to increase the length of some links. When the k th trajectory curve of joint J _k at any time t (t ∈ [0, T]) is j _k (t), the exaggerated value or position e _k (t) of joint J _k is Expressed as

[Equation 1]

e _k (t) = ω _{j k} (t) + (1-ω) a _k (t)

[Equation 2]

a _k (t) = j _k (t) * g _σ (t)

[Equation 3]

Here, a _k (t) is a convolution of j _k (t) and g _σ (t), which means a local spatial average curve of the joint trajectory j _k (t), where g _σ (t) is Gaussian function with standard deviation σ. e _k (t) is an exaggerated value of j _k (t) at a position away from a _k (t). ω is a user defined parameter for exaggerating the joint position, and is a user defined value for dividing a _k (t) and j _k (t) by ω (> 1): 1. The parameters σ and ω are adjustment parameters for determining the degree of exaggeration. a _k (t0) means the average position of the joint at the specified time t0. If the actual position at a certain time t0 is different from the average value, then the distance between j _k (t0) and a _k (t0) will be larger. The interpolated point e _k (t 0) is located away from j _k (t 0). As a result, the curve e _k (t) extrapolated through the motion exaggeration exaggerates the motion by the distance between the actual motion and the average motions.

4 is a reference diagram illustrating a concept of exaggerating a throwing motion using a trajectory process. In Figure 4 the red curve is the original trajectory j _k (t), the green curve is the average curve a _k (t), and the blue curve represents the extrapolated curve e _k (t). As shown in FIG. 4, the motion processed through extrapolation is an unlimited change in the original link-length.

The trajectory-based motion exaggeration method is free from the limitation of link length compared to the existing method of exaggerating the joint angle, thus enabling more dynamic motion exaggeration, and has a simple advantage over the existing exaggeration technique.

In step 106, the motion correction unit 30 modifies the exaggerated motion information generated in step 104. This step is a step of modifying the exaggerated motion in order to prevent the length of the link specified according to the position of the exaggerated joint in step 104 is too exaggerated compared to the original link length.

For example, a, J = {J} _k to J as a set of random is called, all the joints the joint J _k forming the character. Said parent joint (parent joint, J _p) and the child joint (child joint, J _c) the relationship _{(relation) l (J p,} J c) between, and when referred to as a set L of this relationship, L = {l (J _p , J _c )}. The link vector from the parent joint J _p to the child joint J _c can be expressed as v (J _p , J _c ). In this case, for all relationships in the set L, it is necessary to adjust the position of the joint so that the distance between J _p and J _c is closer to the length of v (J _p , J _c ).

5 illustrates an algorithm for modifying exaggerated motion information in step 104. Where T is the number of frames and δ is a predetermined coefficient to prevent distortion of the motion during each iteration. The algorithm is characterized by calculating how much it exceeds the link-length limit, which is an exaggerated motion, and repeatedly correcting the position of the joint. Here m (J _p , J _c ) is a position vector from the original position of the joint to the original position of the parent joint of the joint, and means a position vector corresponding to the corrected position according to the repetition. Each joint moves toward its parent joint, and the overall length of all links becomes shorter with repeated trials. The cyclic loop illustrated in FIG. 5 is repeated until the error is less than the threshold. Here, the threshold value is an intuitive means of adjusting the overall length of the links, and the threshold value can be easily adjusted by adjusting the threshold value.

6 illustrates an example in which an exaggerated motion is modified by repeatedly performing the algorithm of FIG. 5. As the iteration of the algorithm illustrated in FIG. 5 is repeated, the excessively exaggerated link (red) length is modified to have a link (yellow) length within an error tolerance.

In step 108, the joint generating unit 40 generates the joint according to a predetermined criterion based on the motion information input in step 102. In this embodiment, the joint joint layer is introduced to apply the bending action of the rubber rod to the link of the character. The joint generation unit 40 divides the links according to each joint into a plurality of sub-links according to a predetermined criterion, and generates a sub-joint corresponding to each sub link. Position data and orientation data of the joint may be calculated using the original joint position information and orientation information. When the original set of joints is called J , it can be expressed as J = {J _k }. FIG. 7 shows an example of generating a joint structure in step 108. When dividing an arbitrary joint J _k into m _k sub-arthropods, the sub-arthroses can be represented by J ' _{k , 1} , J' _{k , 2} , J ' _{k , 3} , ..., J' _{k , mk} have. Set J _'k = {J ' _{k , l ,} 1 ≤ l ≤ m _k } is a set that consists of all the sub-joints of J _k , and J' = U J ' _k is the sum of all sub-joint sets. In the example of FIG. 7, in the case of J ′ _k , the position of the subjoint is a position that divides J _k and J _{k + 1} by three. In this step, the predetermined criterion is to determine as a joint by the points that divide the distance between J _k and J _{k + 1} by three.

In step 110, the grouping unit 52 determines a rubber-joint group (RJG). The rubber joint group is defined as follows, hereinafter referred to as RJG.

Any end-effector is a member of the RJG.

If the parent joint has only one child joint and the child joint is a member of a certain RJG, then the parent joint is a member of the RJG. If a character does not have fingers or toes, it is possible to group consecutive links of each limb into one RJG. RJGs include, for example, left arm group, right arm group, left leg group, right leg group.

In step 112, the function estimator 54 performs interpolation on the joints having the position specified by the modified motion information in step 106, and estimates a function for connecting the joints to a smooth curve. In this embodiment, interpolation is performed by interpolation for the original positions of all joints for each RJG in order to convert the exaggerated motion into rubber-like motion. For example, the Bezier curve interpolation can be used to estimate the position of the joint. Through this, when the position of each joint is specified, the joint length parameter and the link length between the joints as well as the direction information of the joint can be calculated.

FIG. 8 shows an example of determining the position of the subarticular joint using rubber joint jointing and curve interpolation in steps 110 and 112. The red color represents the original joint, i.e., the exaggerated joint. The position of the original joint is interpolated using a Bezier curve (blue dotted line). The position of the paraarticular treated with a blue circle may be calculated using a curve interpolating the original joint. As shown in FIG. 8, when a blue dotted line, that is, an interpolation curve, is specified, position information and direction information of the joint may be calculated. Interpolation curves are curves that connect joints together and correspond to links. In this step, the criterion for determining the position of the joint is used in the same manner as the criteria used to define the joint in the original joint structure in step 108. You can decide to. Since the joint is above the interpolation curve, the direction of the joint can be determined by calculating the tangential slope at that location.

In step 114, the mass-spring simulation unit 56 performs mass-spring simulation on the motion information modified in step 106. For example, when exaggerating follow-through motion through mass-spring simulations, each joint and sub-joint becomes a particle with mass and a sub-link. Can be treated as a spring. The mass-spring simulation of this step is valid when follow-throw motion is required.

In this embodiment, it is assumed that the follow-throw motion appears for a while after the extreme poses at the beginning and the end of the motion. In this embodiment, the mass-spring simulation is not performed for all the frames, but the mass-spring simulation has a predetermined time from the key frame with respect to the key frames having the extreme poses. Perform up to the dropped frame. Here, the key frame may be selected automatically or manually as a frame effectively indicating the characteristic of the pose. If two frames with extreme poses are selected, the mass-spring simulation is performed from the temporally preceding one of the two frames with extreme poses to the frame located in the middle of the two frames. The mass of the joint and the hook coefficient of the spring can be set singularly. The initial position and velocity of each joint (vaginal point) can be calculated using the positional information according to the RJG generated by the joint motion information generation unit 56, and gravity is modeled as an externally acting force during the simulation.

In step 116, the mixing unit 72 adds the exaggerated motion information for the joints obtained in step 112 and the modified motion information obtained through the mass-spring simulation in step 114 according to a predetermined weight function. If the position of the subjoint is j '(t) and the simulated position is p (t) at any time point t, then j' (t) and p for the mass-spring simulation between times t 0 and t 1 The position b (t) which blended ( t ) is represented by following formula (4).

[Equation 4]

은 시뮬레이션된 모션의 영향을 결정하는 복구 상수(restitution constant)를 나타낸다.

를 크게하거나 작게함으로써 팔로우-스로우 효과의 정도를 초기에 조절할 수 있다.From here,

Denotes a restoration constant that determines the effect of the simulated motion.

You can initially adjust the degree of follow-throw effect by increasing or decreasing.

In step 118, the post processor 74 generates an exaggerated image using the mixed motion information in step 116. Step 118 is largely divided into two steps. The first step is to calculate the orientation information of the joints, and the second step is to generate the skin information. The motion information obtained through the mixing of step 116 includes position information for each joint. Such motion information should be converted into a general skeleton model having position information and direction information. As a method of calculating the direction information according to the joint, for example, the direction of the joints located at the position of the original joint is calculated, and if possible, the directions are adjusted similarly to the direction of the original joint, and then the calculated There is a method of determining the direction of other joints by using slip interpolation between the joints.

On the other hand, skin information is required to generate the final image. In this embodiment, since a new joint structure is generated according to an exaggeration process, skin information used to generate a character using a skeleton model before exaggeration is simply used. There is no number. Therefore, it is desirable to generate the skin information corresponding to the converted joint structure by modifying the original skin information.

Hereinafter, a method of generating skin information suitable for the structure of the subjoint will be described in detail. _Assume that the original joint J _k is assigned to a skin mesh vertex _υ having a weight ω _{υ, k} , and through exaggeration according to the present embodiment, the original joint is connected to the joint J ' _{k , l} (1 ≤ l ≤ m _k ). The original weight ω _{v, k} is redistributed with the new weight ω ' _{v, k, l so} as to be inversely proportional to the distance between the mesh vertex υ and the line segment connecting the joints.

9 illustrates an example of redistributing weights in a cat character in relation to skin information generation. The left side shows the original weights. The right side shows the weight changed according to the structure of the joints generated after the exaggeration. The vertex color of each mesh can be determined by mixing the color of the joint assigned to that vertex with a predetermined weight.

10 is a reference diagram illustrating an example in which a cat character is exaggerated according to the algorithm of the present invention. In FIG. 10, the first row is a character image according to original motion capture information that is not exaggerated. The second row is a character image according to motion capture information exaggerated through RJG positioning without mass-spring simulation. The third row is the character image according to the motion capture information further applying mass-spring simulation and blending. As can be seen through FIG. 10, the character images of the second row and the third row are softly exaggerated images in a rubber form. In particular, the third row of character images makes the resulting motion more dynamic and interesting, and more suitable for cartoon images. In addition, the resulting image due to mixing in the vicinity of the knee may appear undesirably, while this exaggeration can be seen as a kind of "joint fracture" technique.

The time required to perform the above-described algorithm of the present invention for each frame may be calculated through the following experiment. Using a 2.13GHz Intel Core 2 CPU with 2Gb of memory, we calculated the time required for curve extrapolation, mass-spring simulation, blending, and post-processing. millisecond), about 6.22 ms for the character image of the jumping motion, about 13.21 ms for the character image of the kicking motion, and 12.11 ms for the character image of the cossack motion. Since the time is good, it is suitable to be used for exaggeration of real-time video.

On the other hand, the exaggerated image generation method of the present invention can be implemented in a computer-readable code on a computer-readable recording medium. Computer-readable recording media include all types of recording devices that store data that can be read by a computer system.

Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disks, optical data storage devices, and the like, which may be implemented in the form of a carrier wave (for example, transmission over the Internet). Include. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. In addition, functional programs, codes, and code segments for implementing the present invention can be easily inferred by programmers in the art to which the present invention belongs.

So far I looked at the center of the preferred embodiment for the present invention. Those skilled in the art will understand that the present invention can be embodied in a modified form without departing from the essential characteristics of the present invention. Therefore, the disclosed embodiments should be considered in descriptive sense only and not for purposes of limitation. The scope of the present invention is shown not in the above description but in the claims, and all differences within the scope should be construed as being included in the present invention.

Exaggerated image generation method and system of the present invention can freely exaggerate the motion of the character according to the body part compared to the existing motion exaggeration technology, and can produce an exaggerated image having a flexible and dynamic motion bent in the form of rubber It can be used directly in the field of cartoon animation where various expressions are required and character animation in computer games. In addition, the exaggerated image generation method of the present invention can be produced in the form of a plug-in of the 3D graphics program.

1 is a conceptual diagram for explaining an exaggerated image generation algorithm of the present invention.

2 is a block diagram illustrating an exaggerated image generating system according to an embodiment of the present invention.

3 is a flowchart illustrating a method of generating an exaggerated image according to an embodiment of the present invention.

4 is a reference diagram illustrating a concept of exaggerating a throwing motion using a trajectory exaggeration.

5 illustrates an algorithm for modifying exaggerated motion information in step 104.

6 illustrates an example in which an exaggerated motion is modified by repeatedly performing the algorithm of FIG. 5.

FIG. 7 is a reference diagram illustrating an example of generating a joint structure in step 108.

FIG. 8 is a reference diagram illustrating an example of determining the position of the sub-joint using rubber joint jointing and curve interpolation in steps 110 and 112.

9 shows an example of redistributing weights to fit the cat joint structure of a cat character.

10 is a reference diagram illustrating an example in which a cat character is exaggerated according to the algorithm of the present invention.

Claims (17)

In the exaggerated image generation method that exaggerated the motion of the animation character, a) receiving motion information including position information and direction information of a plurality of joints included in an animation character to exaggerate motion; b) generating exaggerated motion information for the plurality of joints by exaggerating the position of the joint based on the trajectory of the joint; c) using the exaggerated motion information and the motion information input in the step a), generating exaggerated motion information on sub-joints positioned between neighboring joints; And d) generating an exaggerated image by using exaggerated motion information on the generated joints. The method of claim 1, Generating modified motion information by modifying the exaggerated motion information in consideration of link lengths between neighboring joints in order to prevent image distortion caused by exaggeration of step b), And using the exaggerated motion information for the joints in step c) uses the modified motion information. The method of claim 2, wherein c) c1) generating a joint located between the plurality of joints input in step a) and positioned between adjacent joints according to a predetermined criterion; And c2) generating exaggerated motion information for the joints using the exaggerated motion information for the joints or the modified motion information and the predetermined criterion used in connection with the joint in step c1). Exaggerated image generation method characterized in that. The method of claim 3, wherein In the case where two joints included in the animation character and adjacent to each other are defined as a parent joint, a joint located relatively inward with respect to the center side of the character, and a joint located relatively outward as a child joint. Grouping joints in which a link between the parent joint having only one child joint and the child joint is continuously formed among the plurality of joints input in step a); And Estimating a function of connecting joints neighboring each other through curve interpolation for joints included in one group by the grouping, And using the exaggerated motion information or the modified motion information for the joints in step c2) uses the estimated function. The method of claim 1, The exaggeration of the position of the joint based on the trajectory of the joint in step b) using curve extrapolation. The method of claim 1, The exaggerated position of the joint based on the trajectory of the joint is an exaggerated image using a convolution of a function representing the position information of the joints based on a time domain and a probability distribution model function representing a probability according to the positions of the joints. How to produce. According to claim 3, following the step c2) And modifying the exaggerated motion information for the sub-joints generated through the step c2) through mass-spring simulation, wherein the mass-spring simulation treats the sub-joints as a quality point and is adjacent to each other. The joints or joints and joints are treated as the force of the spring, In step d), the exaggerated image generating method is generated using the motion information modified according to the mass-spring simulation. The method of claim 7, wherein Step d) mixes the motion information modified according to the mass-spring simulation with the exaggerated motion information for the sub-joints generated in step c) according to a predetermined weight function and uses the mixed motion information. Exaggerating image generation method, characterized in that for generating an exaggerated image. The method of claim 8, The mass-spring simulation is performed on a character belonging to a frame temporally trailing a frame with extreme poses, which is a pose at a time at which the animation character's motion begins or ends, and the weight function The closer the time distance to the frame having a pose is, the greater weight value is assigned to the motion information modified according to the mass-spring simulation. A computer-readable recording medium having recorded thereon a program for executing an exaggerated image generating method according to any one of claims 1 to 9 on a computer. In the exaggerated image generation system that exaggerated the motion of the animation character, An input unit for receiving motion information including position information and direction information of a plurality of joints included in an animation character to exaggerate motion; A motion exaggeration unit configured to generate exaggerated motion information for the plurality of joints through exaggerated position of the joint based on a trajectory of the joint; A limb joint motion exaggeration unit for generating exaggerated motion information on sub-joints positioned between adjacent joints by using the exaggerated motion information and the motion information input through the input unit; And an exaggerated image generating unit configured to generate an exaggerated image using the exaggerated motion information on the generated joints. The method of claim 11, In order to prevent image distortion caused by exaggeration of the position of the joint in the motion exaggeration, a motion for generating modified motion information by modifying exaggerated motion information for the joints in consideration of link lengths between adjacent joints Further includes a correction part, The exaggerated motion exaggeration unit generates the exaggerated motion information further using the modified motion information. The method of claim 12, wherein the ligament motion exaggeration A joint arranging unit defining a joint located between joints adjacent to each other according to the input motion information and a predetermined criterion; And A joint motion information generation unit configured to generate exaggerated motion information for the joints using the exaggerated motion information for the joints or the modified motion information and the predetermined criterion used to define the joint. Exaggerated image generation system, characterized in that. The method of claim 13, In the case where two joints included in the animation character and adjacent to each other are defined as a parent joint, a joint located relatively inward with respect to the center side of the character, and a joint located relatively outward as a child joint. The ligament motion exaggeration A grouping unit for grouping joints in which a link between the parent joint having one child joint and the child joint is continuously formed among a plurality of joints input through the input unit; And a function estimator for estimating a function of connecting joints neighboring each other through curve interpolation for joints included in one group by the grouping. And using the exaggerated motion information or the corrected motion information for the joints using the estimated function. 12. The exaggerated image generating system of claim 11, wherein the exaggeration of the position of the joint based on the trajectory of the joint uses curve extrapolation. The method of claim 11, The mass-spring coil simulation unit may further modify the exaggerated motion information generated by the joint-joint motion information generating unit through mass-spring simulation, and the mass-spring simulation unit treats the joints as a material point and is adjacent to each other. Simulates the joints or joints and joints as if the force of the spring is acting. And the exaggerated image generating unit generates an exaggerated image using the motion information modified according to the mass-spring simulation. The method of claim 16, The exaggerated image generating unit mixes the motion information modified according to the mass-spring simulation with the exaggerated motion information for the sub-joints generated in the sub-joint motion exaggeration according to a predetermined weight function, and mixes the mixed motion information. Exaggerated image generation system, characterized in that for generating an exaggerated image.

Images