KR100903394B1 - Processing Method for Character Animation in Mobile Terminal By Reducing Animation Data - Google Patents

Abstract

The present invention relates to a method for reducing animation data in animating a character using a two-dimensional image in a mobile communication terminal such as a mobile phone and a PDA.

The present invention provides a processing method for character animation in a mobile communication terminal by reducing animation data.

Obtaining still image image data for character animation from a user; Defining an outline of a portion of the still image to be animated; Dividing the portion of the contour to be animated into a triangle mesh; Defining the divided triangular mesh into a shape filled with a bitmap; Moving the frame and converting the characters of the frame to define an animation for each frame; And obtaining the transformation matrix by using the initial coordinates of the triangular mesh in the initial state and the changed coordinates of each frame, and recording the transformation matrix information of each triangular mesh in the animation file.

According to the present invention as described above, while reducing the size of the character animation data in the animation engine, such as a flash engine, character animation and user interface that can achieve the effect that the character set by the user in the two-dimensional image It can be provided.

In addition, the system implemented by the present invention calculates the transformation matrix for applying to the flash engine, and then applies the modified transformation matrix by correcting it only once, so that the gap between the triangle mesh and the mesh is almost invisible when the flash animation is performed. do. In addition, since the number of calculations for finding an optimized conversion matrix can be reduced a lot, it is easy to apply to a mobile communication terminal and a portable device having a small memory and a small calculation capability.

Animated Cartoon, Characters, Mobile Phone, Mobile Phone, Mesh

Description

Processing method for character animation in mobile terminal by reducing animation data

The present invention relates to a processing method for reducing animation data in animating a character using a two-dimensional image in a mobile communication terminal such as a mobile phone or a PDA.

Movies are easy to take due to the high functionality and high capacity of digital devices such as mobile phones, PDAs and digital cameras. However, a video has a much larger capacity than a still picture, and it is very difficult for a user to edit a video captured by the terminal itself.

On the other hand, two-dimensional still image can be easily taken and its capacity is very small compared to video, but it is very difficult to make a still image captured by the mobile terminal itself as a moving image. By displaying the images continuously, it looks like a video.

However, the capacity of displaying a plurality of still images on the screen continuously becomes very large. For example, if you create an animation that is 72 frames (12fps, 6sec) or 24bit color with a 240x320 screen size, you need about 16,588,800 bytes (about 16 megabytes) of memory unless you compress the image. Depending on the compression method, compression requires significant memory.

Therefore, in order to animate all screens into a picture image, a very large memory capacity for storing the animation data of the picture image is required. Therefore, a mobile communication terminal such as a mobile phone or a PDA is difficult to apply due to its small memory size.

As a result, in order to display a video in a mobile communication terminal having a small memory capacity, a special method for dramatically reducing animation data in a mobile communication terminal is needed.

SUMMARY OF THE INVENTION The present invention provides a processing method for character animation of a still image in a mobile communication terminal by dramatically reducing animation data in view of the problems of the prior art as described above.

In order to achieve the above object, the present invention provides a processing method for character animation in a mobile communication terminal by reducing animation data,

Obtaining still image image data for character animation from a user; Defining an outline of a portion of the still image to be animated; Dividing the portion of the contour to be animated into a triangle mesh; Defining the divided triangular mesh into a shape filled with a bitmap; Moving the frame and converting the characters of the frame to define an animation for each frame; And obtaining the transformation matrix by using the initial coordinates of the triangular mesh in the initial state and the changed coordinates of each frame, and recording the transformation matrix information of each triangular mesh in the animation file.

According to the present invention as described above, while reducing the size of the character animation data in the animation engine, such as a flash engine, character animation and user interface that can achieve the effect that the character set by the user in the two-dimensional image It can be provided.

In addition, the system implemented by the present invention calculates the transformation matrix for applying to the flash engine, and if only one correction is applied to apply the modified transformation matrix, the gap between the triangle mesh and the mesh is narrow, and the optimized transformation matrix Since the number of calculations to find the number can be reduced a lot, it is easy to apply to a mobile communication terminal and a portable device having a small memory and a small calculation ability.

According to the present invention, in the case of a character animation that sets a two-dimensional image taken by a mobile phone, a PDA, and a digital camera as a character and animate the character to move, the user of the mobile communication terminal feels a sense of reality rather than watching a still image. We want to reduce the animation data.

Now, the present invention will be described step by step for a case of using a flash engine, which is a representative animation engine.

1. Add character image

The user acquires a still image image through a digital camera, a camera phone, a scanner, or the like, and adds a still image image for character animation on the screen of the mobile communication terminal as shown in FIG. 1. The image may be acquired by itself through the camera module of the terminal, but may be transmitted to the mobile communication terminal from another medium. The transmission method may be various methods such as USB, Bluetooth, IrDA, and the like. At this time, the still image can be any image format supported by Flash (JPEG, PNG, GIF, etc.).

2. Skeleton Insert

Now, the mobile communication terminal inserts the skeleton of the still image image input above from the screen as shown in FIG. Skeletons are not required, but they are added to make the animation more convenient. In the case of standardized characters, the animation can be defined a little more easily by inserting a predefined skeleton. The inserted armature allows the user to adjust the movement and zooming by using input devices such as a touch screen and a stylus pen.

Object skeletons, street lamps, object skeletons such as trees, and pet skeletons such as dogs and cats can be variously defined and shared by users.

3. Contour Definition

Now, as shown in FIG. 3, only the portion of the still image region added previously that is to be animated is defined as the outline of the character (person, tree, pet, etc.).

To smooth the animation, the joints of the character define the contours a bit more precisely.

Depending on the sophistication of the contours, the triangular mesh discussed in the next step will be refined and the animation smoothed.

4. Split triangular mesh inside contour

Using the skeleton and the character outline information, the inside of the outline is divided into triangular small pieces (Mesh, hereinafter referred to as a mesh) as shown in FIG. Skeletal information is not necessary when dividing into meshes, but having a skeleton makes the joints clearer, affecting the quality of the animation, making the mesh near the joints a little smaller, and making the other parts a bit larger. The number of can be reduced somewhat. This method is cumbersome to set the skeleton for the character, but it is efficient. If you split only a contour into a triangular mesh, you should minimize the size of the mesh as much as possible to smooth the animation.

The speed of dividing into a triangular mesh is also important when dividing into a mesh, but the mesh size and placement is more important because the animation motion is smooth when the mesh and mesh are properly sized and placed properly.

If you don't add a skeleton or don't have a skeleton for your character, you can get the same effect as adding a skeleton without adding a skeleton by just setting the joint position before creating the mesh.

5. Define a segmented triangular mesh as a shape filled with a bitmap

Using the divided mesh and the image as shown in Figure 4, it is necessary to define a shape (shape) that can be displayed in the flash.

It is possible to add an image of the entire input character to make a triangular piece of image for every mesh, but in this case there will be no increase in the total image capacity, but the process of dividing into a piece of image is very time consuming. Therefore, in the present invention, the image of the entire input character is used as it is without conversion. To do this, two steps are needed.

The first step is to define the image of the entire character, and the second step is then to define all the triangle meshes, giving them a bitmap-filled shape with all the mesh images.

The process of defining a character image is defined using flash image definition tags (DefineBits, DefineBitsJPEG2, DefineBitsJPEG3, DefineBitsLossless, DefineBitsLossless2, etc.). Detailed descriptions of the tags (commands) of such a flash engine are well known to those skilled in the art, so the following descriptions are omitted. For details, refer to the documentation on the Flash file format of Adobe (Macro Media Co., Ltd.) You can use a web search engine such as Google, Google, or Naver, or visit the website ( http://www.the-labs.com/MacromediaFlash/SWF-Spec/SWFfilereference.html ).

The process of defining the triangle is defined using the flash shape definition tags (DefineShape, DefineShape2, DefineShape3, etc.). When defining, define the shape by the triangle image, not the square image, and set the bitmap-filled attribute.

After these two processes, the number of triangles is somewhat increased, so the capacity is somewhat increased, but the increase in capacity is not so large. Since triangle meshes are defined only in the initial frame, the number of triangle meshes does not increase as the frame increases.

Due to the large number of meshes, there are many shape definition tags, but since the binary image is added only once, the size increase is not significant.

And even if the entire image is divided into triangles, there is no difference in the image quality seen in the actual flash. In terms of display speed, a difference may occur depending on the number of meshes, but the difference in display speed is not so large in a mobile communication terminal.

Through this process, the actual file of the flash is recorded as shown in FIG. First, the basic header information of the flash is recorded, and image data (IMG1) is written, and information of triangle meshes having attributes filled with the image data is recorded. After all this information is recorded, the first frame is finished by using a ShowFrame tag.

6. Define animation by frame

As shown in Figure 6, while moving the frame, the animation is defined by converting the characters of the frame. You can define the animation by moving the armature or by moving and transforming the shape of the outline.

7. Record Animation to Flash File

Now record the animation information in the flash file. In the recording method, image data and a triangular mesh (hereinafter referred to as shapes) filled with the image data are first recorded as shown in FIG. 7 below. If there are additional shapes other than this, such as a background image, they are recorded together, and a frame end tag (ShowFrame tag) is recorded to end the initial frame.

Then, after moving to the next frame, a transformation matrix is obtained using the initial coordinates and the current coordinates of the triangular mesh using the movement information of the handle of the frame for each frame and the information of each mesh. Using a tag (PlaceObject, PlaceObject2 tag, etc.) defining a shape transformation, the transformation operation of each triangular mesh is recorded, and then the frame end tag (ShowFrame tag) is recorded to end the corresponding frame.

In this way, after recording all the frames, the file is terminated.

The structure in which this information is written to the file can be simplified as shown in FIG.

The transformation matrix required to define the transformation of the triangle mesh is, when converted from triangle T0 to triangle T as shown in Fig. 8, mathematically there is one 3x3 transformation matrix. 6 factors (X / Y Scale, 2 Rotation & Skew factor, X / Y transformation) of transformation matrix that define transformation of 2D triangle among tags of flash (PlaceObject, PlaceObject2 tag) Etc.) to record the conversion operation information in the file.

7.1. Mathematical Transformation Matrix and Flash Transformation Matrix Difference

Mathematically, the meaning of the six elements of the matrix needed to define the transformation of triangles is the same in Flash. However, in Flash, the six factors of the 3x3 transformation matrix equation 1 (Sx: Scale X, Sy: Scale Y, R0: 1st rotate and skew, R1: 2nd rotate and skew, Tx: Translate X, Ty: Translate Y Are not all real numbers.

Scale X / Y, 1st & 2nd rotate & skew values are 16bit fixed-point real numbers, and Translate X / Y is treated as Twips (1440 TWIPS = 1 inch & 72 points = 1 inch). Thus, the actual Tx and Ty arguments cannot be used as they are, they are multiplied by 20, and converted into integers and stored. That is, in Flash, the Translate X / Y factor is truncated to 1/20 (0.5) units or less, so that a large error occurs.

Due to this difference, applying a mathematically obtained transformation matrix directly causes a problem of gaps between meshes in actual flash.

After getting the transform matrix, if the value of transform x / y is converted to Twips among the matrix values, fortunately it is only an integer, there will be no error, but if there is a value below the decimal point, there will be some error. If a shape is only zoomed in and moved, it can only appear as an integer in the screen coordinates. However, if the rotation and distortion are performed, the translate x / y factor of the transformation matrix is also changed, and only integers cannot be obtained.

Essentially, if you change the file format of Flash so that Adobe uses the translate x / y argument by mistake, this problem will go away.

In the present invention, a method for minimizing a gap between a mesh and a mesh that can be used even if Adobe translates the translate x / y factor in the future will be presented in the next chapter.

7.2. Find an approximation of the transformation matrix

As shown in Figure 9 below, one triangular mesh in the initial state is called T0, and the converted triangular mesh of T0 in the current frame is called T.

As shown in Fig. 9, since T0 (Equation 2) and T (Equation 3) are known, a 3x3 transformation matrix M (Equation 4) satisfying T = T0xM can be obtained.

(Initial coordinates of the triangle)

(Triangle coordinates of the current frame, target triangle matrix)

Since T = T0xM, the transformation matrix M is M = T0-1xT.

Type casting is performed for each of the six parameters in units of M used in flash. A matrix R of 3 × 3 is obtained by performing matrix multiplication of the initial mesh T and the matrix M1 (type converted form of M). It is checked whether the matrix R thus obtained coincides with T, which is mesh information of the current frame.

When checking for a match, a normal case does not match as shown in FIG.

If the two meshes do not match, as shown in Figure 10, there are many ways to find the optimized solution, the present invention proposes the following method (see Figure 11).

Find one point C inside T.

2. Find the distance from C to the corners of T and R (dT1, dT2, dT3, dR1, dR2, dR3).

3. Find the difference between the distances between the edges of T and R (d1 = dT1-dR1, d2 = dT2-dR2, d3 = dT3-dR3).

4. Since negative edges among d1, d2, and d3 as shown in FIG. 10 are inward, the corresponding edges are moved outward by a certain distance (| distance difference | * 1.1) in the direction of the corresponding corners of C to T. Since the positive edges of d1, d2, and d3 are outward, the corresponding edges are moved inward by a predetermined distance (| distance difference | * 0.3). (Where | distance difference |: absolute value of the distance difference)

5. With the modified T, perform the process of Figure 9 again until you find the optimized solution.

The above method is not intended to match the target mesh T and the mesh R to which the transformation matrix is applied, but to make the mesh and the mesh overlap slightly by making R larger than T. As shown in Fig. 11, since the angles of the corners of T and R from the inner one point C do not coincide exactly, the method of the present invention cannot be mathematically verified, but the deviation of the angle is small, and also between the mesh and the mesh. In order to eliminate the gap, the proposed method is to minimize the overlap between meshes and reduce the number of calculations.

Instead of the above method, if you obtain a matrix by increasing the three corners of the triangle in small increments at a point C inside the triangle, the speed is very slow, and you cannot determine how much you need to increase or decrease the number, and you can't specify the number clearly. There is.

For example, even if you want to check only 10 steps up to 0.1 by 1, it is necessary to perform about 3 edges, so it calculates 1000 times, 103 times, which is very inefficient. However, this method is not optimized if the transformation matrix is corrected only once in each corner, but is not optimized. However, triangles and triangles overlap to some extent, so that triangles without spacing between two triangles are obtained. You need to verify while optimizing, but you only need to do it once or twice to improve speed.

8. Overall Configuration and Flowchart of the Invention

To generate a flash file for character animation using the raster image, the procedure is as shown in FIG.

1. First, add a raster image obtained using a digital device (S1).

2. Insert a skeleton that fits the added image (S2). Skeleton can be not only human, but also animals and plants, and the insertion of the skeleton is for convenience of animation definition.

3. Define the outline (outline) of the character (S3). Although the contour of the character can be automatically recognized, the terminal is defined by the user in consideration of the processing speed of the central processing unit. (It can also be recognized automatically.)

4. Automatically or manually decompose the specified contour into a triangle mesh (S4). When the skeleton is inserted, the division of the mesh is not evenly divided, but the fracture portion is more precisely divided, and the portion that is not, the mesh can be largely divided.

5. The user defines an animation while moving the frame (S5).

6. Create a flash file to be created (S6). (Header information such as file version and frame number is recorded at the beginning after file creation.)

7. Record image data in the generated file (S7).

8. Define as many triangle shapes as the number of divided meshes (S8). At this time, all shapes are defined as bitmap filled shapes added in step 7.

9. End the initial frame and move to the next frame (S9). Write the end frame tag to a file.

Now, to move the frame and write the conversion information of the shape defined in step 8 to the file, repeat 10 to 15 times below the number of frames.

10. First, all meshes are generated from the defined animation information of the corresponding frame (S10).

For every mesh, repeat steps 11 through 14 for the number of meshes to find the transformation matrix for each mesh.

11. Using the mesh generated in step 4 (MESH0 []) and the mesh generated in step 10, a transformation matrix is obtained (S11).

12. The transform matrix M is corrected to obtain a modified transform matrix M1 (S12). You can use the method described in 7.2. Calibration is a necessary step to reduce the gap between meshes due to the limitations of the current flash file and can be skipped. Also, if the file format of the flash in the future changes all the arguments of the conversion matrix to a real number, it may not be necessary. have.

13. Record the conversion information of the shape defined in step 8 (S8) to a file. Record using the ID of the shape and the transformation matrix found in 12 (if skipped 12, use the transformation matrix 11).

14. After repeating as many as 11 to 13 times the number of meshes, the frame is terminated, and moves to the next frame (S14). Write the end frame tag to a file.

15. After repeating the number of frames from 10 to 14, finish writing the file.

1 shows a still image image obtained for the present invention.

Figure 2 shows a still image image with the skeleton inserted for the present invention.

Figure 3 shows a still image image with the outline inserted for the present invention.

Figure 4 shows a still image image divided into triangular mesh for the present invention.

5 shows the physical file structure of a flash used in the present invention.

Figure 6 illustrates defining frame-by-frame animation for the present invention.

Figure 7 shows the structure of the animation data recorded in the flash file for the present invention.

8 shows the relationship of the transformation matrix used in the present invention.

9 shows a process for obtaining a transform matrix in the present invention.

FIG. 10 shows that the target mesh T does not coincide with the mesh R to which the transformation matrix is applied in the present invention.

11 shows a process of obtaining a modified target mesh T in the present invention.

12 is an overall flow chart of the present invention.

Claims (9)

In the processing method for reducing the animation data to animate the character in the mobile communication terminal, Obtaining still image image data for character animation from a user; Defining an outline of a portion of the still image to be animated; Dividing the portion of the contour to be animated into a triangle mesh; Defining the divided triangular mesh into a shape filled with a bitmap; Moving the frame and converting the characters of the frame to define an animation for each frame; Obtaining a transformation matrix by using the initial coordinates of the triangular mesh in the initial state and the changed coordinates of each frame, and recording the transformation matrix information of each triangular mesh in an animation file, The animation uses a flash animation engine, The transformation matrix is a transformation matrix after the process of minimizing the spacing between the meshes generated when the mathematically derived transformation matrix is applied to the flash animation engine. Processing method for character animation. The method of claim 1, And inserting a skeleton of the obtained still image to reduce animation data and to animate a character in the mobile communication terminal. delete delete The method of claim 1, The minimization process is a process for making the animation data characterized in that the meshes overlap a little by little, the processing method for character animation in the mobile communication terminal. The method of claim 5, The overlapping process, Finding a point C of the target mesh T; Obtaining first distances (dT1, dT2, dT3, dR1, dR2, dR3) from each point (C) to each edge of the target mesh (T) and the mesh (R) to which the transformation matrix is applied; Differences (d1 = dT1-dR1, d2 = dT2-dR2, d3 = dT3-dR3) to each corner of the target mesh T and the mesh R to which the transformation matrix is applied using the first distances Obtaining a; Since the negative edges of the differences d1, d2, and d3 are inwards, the corresponding edges are moved outwardly by a predetermined distance in the direction of the corresponding edge of the target mesh T from a point C. Since the positive edges of the differences d1, d2, and d3 are outward, the character animation in the mobile communication terminal is reduced by reducing the animation data, which is performed by moving the corresponding corners inward. Treatment way for you. The method of claim 1, And the mobile communication terminal is a mobile phone to reduce the animation data. The method of claim 1, And the mobile communication terminal is a camera phone, thereby reducing animation data. The method of claim 1, And the mobile communication terminal is a PDA.

Images