JP3467406B2 - Animation generation method and computer graphics - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an animation created based on motion data, and more particularly, to an animation generation method capable of various and autonomous modification of motion data applied to a model, and a computer graphic thereof. .

[0002]

2. Description of the Related Art Key frame animation is well known as a method for generating a three-dimensional animation, but it is difficult for this animation to represent a natural motion of a human being or the like. Therefore, there is a method of motion capture in which an actual operation is imaged and reproduced. However, since the amount of data in motion capture is large, it is difficult to edit motion data and create a new motion.

As one of the solutions to this problem, Japanese Unexamined Patent Publication No.
There is a "motion expression method for an articulated object" proposed in No. 71078. Here, animation is generated by converting motion-captured motion data into a Fourier series (FFT parameter) and interpolating a plurality of motion data. In addition, if the operator specifies a "adverb" that modifies the basic motion, for example, "walk" of the basic motion
It is possible to express the motion of "walking happily" by modifying the "adverb" with "fun".

[0004]

SUMMARY OF THE INVENTION The method of the above cited example is a motion capture method such as interpolating between a plurality of motion-captured basic motions or generating a change motion of a designated "adverb" from the basic motions. It becomes easy to generate a new operation based on the obtained data.

However, in order to modify the basic movement, "adverb"
The modification operation of the frame after specifying is fixed. In addition, if a number of models are modified with the same "adverb", the modification action will be the same even if the models are different, resulting in an unnatural animation as a whole. Therefore, in order to generate a more natural animation, it is necessary to frequently specify the "adverb" for changing the motion for each frame and for each model, resulting in a huge amount of operator work. There is.

It is an object of the present invention to provide an animation generation method and a computer graphic capable of overcoming the above-mentioned problems of the prior art and easily producing a variety of natural animations.

[0007]

According to the present invention for solving the above-mentioned problems, a motion database storing a plurality of periodic motion data of a plurality of models and characteristic motion data following the periodic motion. Therefore, in the method of generating an animation consisting of a plurality of frames by designating a desired motion, path (path) and model, the motion name, the model characteristic consisting of the characteristic characteristic and the ability characteristic of the model, and the number of frames are designated. The first adapted to refer to the operation schedule and reflect the specified operation name and the specified operation characteristic from the operation database, the optimum cyclic operation data and the specified ability characteristic.
Characteristic data is selected, the former motion data is reflected in all frames, and the latter motion data is changed by evaluating the change in the performance characteristics for each frame. To synthesize.

That is, the characteristic of the model is set with the characteristic such as the physical strength as well as the specific characteristic such as sex and age, the former is used for the selection of the optimum periodic motion data, and the latter is used for the selected basic motion. On the other hand, when synthesizing a modification action such as “tiredness”, a decrease in physical strength is calculated for each frame, and the synthesis ratio is changed according to the result.

Similarly, when the environmental characteristics such as the temperature of the virtual space in which the model is present are set and the modifying action such as "cold" is synthesized, the change of the temperature is calculated for each frame, and according to the result, Change the composition ratio.

Further, according to the present invention, the periodic motion data and the stochastic motion data following the periodic motion are accumulated, and the optimum cyclic motion data for the designated motion name and model characteristic is stored from the motion database. It is selected and reflected in all frames, and whether or not a probabilistic operation is performed is determined from the specified frequency for each frame, and if it is performed, the probabilistic operation data of the specified operation name is added to the periodic operation data of the frame. To synthesize.

The periodic motion data is "run", "walk", "stop", etc., and the stochastic motion data is "raise hands", "shake head", etc., and one of the former motions in the motion schedule. The latter action name and frequency are set for the name. For example, if the frequency is 50%, a random number is generated every two frames, and it is determined from the result whether or not the stochastic motion is combined. According to this, even if the same probabilistic operation is set for many models on the frame screen, the execution of the operation becomes probabilistic and is not uniformized.

Further, according to the present invention, periodic motion data and a plurality of motion transition sequences indicating the transition order of the motions are stored in the motion database, and the optimum cyclic motion data is stored together with the optimum motion from the motion database. When the leading motion transition sequence is selected, a frame is created starting from the selected periodic motion data, and when the change in the performance characteristics is evaluated for each frame and the change exceeds a threshold value,
A frame is created by changing to the periodic motion data of the next motion name in the selected motion transition sequence. Alternatively, environmental characteristics are specified in the operation schedule, and when the characteristic characteristics of the model and / or the environmental characteristics change every frame and exceed a threshold value, the periodic operation data is changed.

According to this, the basic scheduled operation itself can be changed. For example, it can be changed from "run" to "walk" when the physical strength is below a predetermined value, or can be changed from "walk" to "run" when "rain" begins to fall due to environmental characteristics.

The computer graphics of the present invention is
In addition to having a function for realizing the above-described animation generation method, a user interface for displaying the operation schedule input screen and the menu screen corresponding to each setting content to improve the user's input is provided.

[0015]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 shows a functional block diagram of an animation creating apparatus using computer graphics.

The animation creating apparatus has a database 1, an operation schedule table 6 for storing operation schedules, a stochastic operation table 7 for storing stochastic synthesized operations, a model characteristic table 8 for storing model characteristics, and a model. An environment state table 9 that stores environment characteristics of the virtual space, a user interface 10 that stores information in tables 6 to 9, and optimum operation data to be applied to the model is selected from the operation database 21 based on the setting contents of the tables 6 to 9. A motion selecting unit 2, a composite motion generating unit 3 for creating motion data of each joint of a model by synthesizing the selected motion data and feature amount data,
It comprises a movement amount generating means 4 for generating a movement amount along a path based on motion data, and a display means 5 for displaying an animation.

The database 1 includes a motion database 21 in which models, joint structures, and actual data are hierarchically structured for each motion category, a model characteristic database 22 that stores model characteristics, and a joint structure database 23 that stores model joint structures. , A motion transition database 24 that stores a combination of a plurality of motion transitions (motion transition sequence).

FIG. 2 shows the configuration of the operation database according to one embodiment. The motion database 21 is extracted from the cyclic motion data part 211 and cyclic motion part 211 that retains data for one cycle for each basic motion such as “walking”, and the basic motion is “energetic”, “tired”, etc. Of the feature amount data section 212 that holds the motion data modified according to the “adverb”, and holds the motion data such as “raise your hand” and “fall” as raw data (FFT conversion is not performed as it is in motion capture). RAW operation data section 213,
It has a motion transition table 214 that stores combinations of motions that change according to transition conditions such as “run” → “walk” → “stop”. Each motion data section is composed of a motion category 215, model characteristics 216 that perform a motion, a joint structure 217 of motion data, and actual motion data 218. The actual operation data 218 is FFT parameter or RAW motion data.

The structure of the motion database 21 shown in FIG. 2 has a model characteristic DB 22 and a joint structure DB included in the database 1.
23, the action transition DB 24 is edited for each action category, and actually stores the connection relation with the data of other databases. Therefore, the configuration of other databases is also obvious from FIG. 2, and detailed description thereof will be omitted. The operation transition table 21 of FIG.
Reference numeral 4 is the motion transition database 24 itself, which actually has a different configuration from the motion database 21.

FIG. 3 shows the joint structure of the model.
The joints that make up the model are hierarchized in a parent-child relationship, and the position between joints and the angle information of each joint are managed as relative values from the parent joint. For example, if the position information of the waist joint 31 that is the highest parent is changed, the positions of the neck joint 32 and the hip joint 33 that are the children of the joint 31 can also be changed in the virtual space. Similarly, the positions of all joints such as the joints on and after the joint 34 on the left shoulder, which is a child of the joint 32 on the neck, can be changed.

FIG. 4 is an explanatory view of the joint structure of the model and the FFT motion data. Cyclic motion data 1
In the eleventh example, the motion data type 41, the total number of joints 42, the joint position information 43, the number of frames in one cycle 44, and the angle information of each joint are FFT parameterized actual data 45. When the number of frames in one cycle is 20, there are 20 rows of actual data. Each line is x in the order of the parent and child of the joint.
The FFT parameterized angle information in the axis, y-axis, and z-axis directions alternately describes DC data indicating the initial position of operation and AC data indicating the changed position.

FIG. 5 shows an explanatory diagram of RAW motion data. In the example of the RAW motion data 113, the motion data type 51, the total number of joints 52, and the joint position information 5 are the same as in FIG.
3, the number of frames 54 of RAW operation data and the actual data 55 are described. The difference from the cyclic motion data is that the actual data 55 becomes raw angle information.

FIG. 6 shows the structure of the user interface. The user interface 10 is an input / output unit configured on the screen, and an operation applied to the model (for example,
A motion selection button 61 for selecting "walk") and its display area 62, a display area 63 for inputting the number of frames in one cycle of motion, and a motion path selection button 64 for selecting a motion path (for example, "path A"). And its display area 65, the stochastic operation setting button 66 and the identification number display area 67 in the set stochastic operation table 7, and the model characteristic setting button 68 and the identification number display area 6 in the set model characteristic table 8.
9, an environmental status setting button 610, an identification number display area 611 in the set environmental status table 9, and a schedule display area 6 for displaying a list of a series of schedule setting statuses
15 and a total frame display area 616 are provided.
Other configurations of the user interface 10 are as follows.
It will become clear in the following description.

Next, an animation generation operation according to this embodiment will be described. First, the user selects interface 1
Creation of an operation schedule performed by setting 0 and setting of tables 6 to 9 will be described. When the selection button or the setting button described above is turned on (pressed), a menu processing function (not shown) built in the user interface 10 is activated and a corresponding menu screen is displayed.

FIG. 7 shows an example of the operation selection menu.
When the user selects an arbitrary action from the action menu screen, the user interface 10 displays the action in the selected action display area 62.

FIG. 8 shows an example of the path selection menu.
Paths A to C are paths for which route coordinates are set, and objects A and B are paths for which destination coordinates are set. The operation when stopped is “no path”. An identification number is assigned to each path, and the identification number of the path selected by the user is displayed in the operation path display area 65.

FIG. 9 shows an explanatory diagram of the operation path. In the virtual space, a human model 91, a curve 92 having a start point 93 and an end point 94, and an object 95 exist. Curve 92
A name is given to the cube 95. A curved line 92 is a fixed path with route coordinates set. When the object 95 is designated as the motion path, the human model 91 performs the designated motion assuming a virtual path 96 that goes straight from the current position in the virtual space toward the object 95. In addition, without setting a specific object, for example, it is possible to assume the virtual path 97 in the direction in which the human model 91 is currently facing and perform the designated operation.

FIG. 10 shows an example of the stochastic operation setting menu. When the stochastic operation setting button 66 is pressed, the establishment operation setting menu screen is displayed. On this screen, the registration number is entered in the probabilistic motion display area 101. Next, a person who performs a probabilistic operation, for example, “waving hands” is selected from the operation selection area 102, and the frequency of executing the operation, for example, “30%” is input to the frequency input display area 103. When the registration button 105 is pressed, the selected stochastic motion and its frequency are displayed in the stochastic motion list display area 104.

When setting a plurality of stochastic operations, the above procedure is repeated. When changing the contents of the stochastic operation list display area 104, the corresponding line is designated and the delete button 106
Press to reset after deleting the contents. Finish button 107
When is pressed, the contents displayed in the stochastic motion display area 104 are stored in the stochastic motion table 7. After setting the probabilistic operation in this way, the user selects the area 67 on the input screen of FIG.
Enter the registration number in and set the probabilistic action to use for the action schedule.

FIG. 11 shows an example of the model characteristic setting menu. When the model characteristic setting button 68 is pressed, the model characteristic setting menu screen is displayed. Here, the characteristic number is input in the identification number display area 111, and the model characteristic is set in the model characteristic input area 112. Type 113, gender 114, age 115, physical fitness 116, physical condition 117, etc. can be input as model characteristics. The physical strength is arbitrarily scored, for example, with a maximum of 10 points. When the setting end button 118 is pressed, the input information is stored in the model characteristic table 8. After setting the model characteristic, the user inputs the characteristic number in the area 69 of the input screen to set the model characteristic used for the operation schedule.

FIG. 12 shows an example of the environmental condition setting menu. When the environmental condition setting button 610 is pressed, the environmental condition setting menu screen is displayed. Here, the environment number is input in the identification number display area 121, and the environment state input area 11
Set the environmental condition to 2. Environmental temperature of 123,
Light intensity 124, wind power (wind speed) 125, etc. can be input. When the end button 126 is pressed, the input environmental information is stored in the environmental status table 9. After setting the environmental state, the user inputs the environmental number in the area 611 of the input screen to set the environmental characteristics used for the operation schedule.

The operation and path selected by the series of menus and the identification information of the setting tables 7 to 9 are displayed on the input screen of the user interface 10 together with the number of frames set for each operation schedule. In this state, when the schedule registration button 612 is pressed, the schedule display area 615 displays an action schedule including actions, the number of frames, action paths, stochastic actions, model characteristics, and environmental conditions.

The user repeats the above input and displays a plurality of operation schedules in the schedule display area 615. To change the displayed schedule, select the schedule line to be changed, change the information, and press the change schedule button 613. When deleting the displayed schedule, the line of the schedule to be deleted is selected and the schedule deletion button 614 is pressed. The total frame number of all schedules currently displayed is displayed in the total frame display area 616.

The displayed operation schedule is stored in the operation schedule table 6 when the animation generation button 617 is pressed. If you press the end button 618,
The information stored in each table is discarded and the process ends without generating an animation.

Next, a processing procedure for creating an animation will be described. When the setting of the schedule table 6 is completed on the user interface 10 and the animation generation button 617 is pressed, the operation selection unit 2 is activated.

FIG. 13 shows a processing flow of the operation selecting means. The motion selecting means 2 first refers to the motion schedule table 6 and checks whether the animation to be created is the final frame (s10). If it is not the final frame, the process from the operation data selection process (s11) to the operation transition selection process (s14) is executed, the next-stage synthesized action generation means is activated (s15), and the process for one frame is completed.

FIG. 14 shows the operation data selection process (s11).
Shows the details of. First, it is determined whether the current frame is a new operation schedule in the operation schedule table 6 (s111). If it is a new schedule, the operation name and model characteristic number designated for the schedule are acquired from the table 6 (s112), and the corresponding model information is acquired from the model characteristic table 8 (s113). And
The optimum motion data for the specified motion and model is searched from the motion database 21 (s114).

At this time, it is checked whether there is data in which the specified action and the model completely match (s115), and if there is completely matching action data, it is selected (s11).
6). If there is no motion data that completely matches, the motion data that is closest to the specified model characteristics among the searched data is selected (s117).

For example, when the action specified by the schedule is "walking" and the model characteristics in that frame are "human", "male", and "25 years old", the action selecting means 2
Searches the motion database 21 and attempts to select motion data whose category is “walking” and whose motion characteristics match the model characteristics. Then, when there is no completely matching data, the specified behaviors are matched, and the model characteristics “human”, “male”, and “25 years old” are searched for in the order of closeness. Normally, for "walking" motion data, "human",
Since “male” is prepared, the motion data whose age is closest to 25 is selected.

Next, the motion selection means 2 performs a probabilistic motion selection process (s12). FIG. 15 shows details of the probabilistic motion selection processing. First, it is checked whether a probabilistic motion has been selected for the current motion (s121). If the stochastic operation has already been selected, this processing ends.

When the probabilistic action is not selected, the following process is repeated until the process for all registered data is completed or one probabilistic action is determined. Therefore, the probabilistic motion table 7 is referenced by the registration number of the motion schedule, the total number of set probabilistic motion data is acquired, and the following process is repeated until the number of registered data is exceeded (s122).

The action name and frequency of the probabilistic action data are acquired (s123), and it is determined whether or not a random number is generated this time from this frequency (s124). For example, if the frequency is 50%, a random number is generated every two frames. When a random number is generated (s125), it is checked whether the random number is larger than a specific value (s126). In this example, it is determined whether the random number is larger than 0.5, and if it is larger, a probabilistic motion is added to the current motion (s127). When the probabilistic action is determined, this processing is immediately terminated.

For example, registration 1 set to "run (75 frames, path C)" in the operation schedule of FIG.
The probabilistic motion of “waving hands (frequency 30%)” in FIG.
And "Shake your head (frequency 50%)". In this case, whether or not to add a probabilistic motion of "waving" is evaluated by generating a random number for every 30% of 75 frames of "running", that is, every 3 frames. Similarly, the "shaking" motion is performed by generating a random number every two frames for evaluation.
When the random number exceeds 0.5, for example, the stochastic operation is performed in the frame.

The stochastic operation of this embodiment is executed until the end of the basic operation, and while one stochastic operation is being executed, evaluation of another stochastic operation is stopped. That is, there is only one stochastic operation in the unit operation schedule. This is because if a plurality of stochastic actions are simultaneously established, it becomes unnatural as a human action. Therefore, when a plurality of stochastic actions are set in one action schedule, both the selection of the action and the necessity of execution are determined by random number generation.

Next, the operation selecting means 2 performs a selecting operation of the combining operation according to the environmental condition (s13). FIG. 16 shows details of the combining operation selection process depending on the environmental state. Here, the environmental characteristics (temperature, luminous intensity, wind power) of FIG. 12 will be described as an example.

First, referring to the environmental condition table 8 from the environmental number in the schedule, it is determined whether the set temperature is below a specific value (here, 10 degrees) (s131). When it is equal to or more than the specific value, the synthesis ratio of the feature amount data of "cold" is set to 0 (s132). If the temperature is below a certain value,
The synthesis ratio of “cold” is obtained by, for example, synthesis ratio = 10−air temperature + 100 (%) (s133).

This "cold synthesis ratio" is used for each FFT described later.
In the parameter combining process, the “cold” FFT data is multiplied. Incidentally, the “cold” FFT data stored in the feature amount data section 212 is motion data that causes the body to “smoothly” shake, and is added to the basic motion by increasing / decreasing the data at a synthesis ratio. "Cold synthesis ratio"
May be obtained as a value corresponding to a step change in temperature.

Next, it is determined whether the set luminous intensity value is "night" (s134). If not "night",
“No deviation” is set for the operation path (s135).
In the case of “night”, when the model moves along the curved path while correcting the direction, it is dark, so it is assumed that an error will occur in the angle of the direction to be corrected. On the other hand, for example, "+ 20% shift" is set (s136).

Next, it is judged whether or not the wind force against the traveling direction is a specific value (for example, 10 m / s) or more (s13).
7). When the wind force is smaller than the specific value, the movement amount is not subtracted from the progress of the model (s138). When the wind power is above a specific value, the moving amount subtraction formula (for example, 10
(Wind force-10) (%)) is set to the subtraction ratio of the movement amount according to the wind force (s139). By the way, wind force 15m /
At s, the subtraction ratio is 50%, and the model advances by half.

Next, the action selecting means 2 performs the action transition selecting process based on the evaluation of the model characteristic (s14). In the present embodiment, when the physical strength of the model decreases due to the progress of the action schedule and the transition condition is satisfied, the action transition according to the action transition sequence is performed.

FIG. 17 shows a flow chart of the operation transition selection processing. First, a physical strength value is acquired from the model characteristic table 8 (s140), and the physical strength value is subtracted (here, 0.01).
Is performed (s141). Next, it is determined whether the action transition sequence to be applied to the designated action has been determined (s142). If not decided, the action transition table 21 using the designated action as a key
4 and searches for a corresponding motion transition sequence (in FIG. 2,
Obtained as "run" motion transition sequence (3) (s1
43), a random number within the number of candidates is generated to determine the motion transition sequence to be used (s144).

Next, the physical strength of the model is a specific value (here,
3) It is determined whether or not the above (s145). When the physical strength is equal to or greater than the specific value, the first motion in the motion transition sequence, which is the designated motion, remains as it is, and the motion transition is not performed. However, a composite ratio of "fatigue" of the model, for example, a composite ratio that increases by 0.1% is calculated (s149), the current physical strength value is stored in the model characteristic table 8 (s1491), and the process ends. In addition,
The motion of "tiredness" is reflected in, for example, a forward leaning posture and a decrease in arm position according to the composition ratio.

On the other hand, the physical strength of the model is 0.
Since 0 is subtracted, the physical strength that was 5.0 at first becomes 3.0 at the 200th frame (after about 10 seconds). Then, if the physical strength of the model is less than 3.0, it is determined whether the physical strength is 0 to 3 (s146), and if so, the schedule is changed so that the second motion of the determined motion transition sequence is performed ( s147). Further, when the physical strength is 0 or less, it is determined whether or not the third movement exists in the movement transition sequence (s148), and if there is the third movement, the schedule is changed to the third movement (s149). The second movement,
Also in the case of transition to the third action, the "fatigue" composite ratio is calculated.

As a result, the operation schedule is changed from "run" to "walk" in accordance with the decrease in physical strength of the model.
In addition, you can transition to the third stop. Further, since the motion transition sequence is selected for each model by a random number, even when a plurality of models “run”, the motion and timing of the next transition of each model are not standardized.

In the above, the transition of the movement depends on the physical strength of the model, but it is also possible to evaluate the deterioration of the physical condition by an exponent or a random number and make a transition to a movement such as "squatting". Further, various variations are possible such as "rain" is caused to fall in the virtual space due to environmental characteristics, and "walking" is changed to "running" depending on the amount of rain.

By the series of processes described above, the motion data and the composition ratio for composing the motion of one frame are all determined. Next, the action selecting means 2 activates the synthesizing action generating means 3 (s15), and then returns to the process s10 to proceed to the action data selecting process of the next frame. On the other hand, the activated composite motion generating means 3 executes an inverse FFT operation using the motion data and the composite ratio determined by the motion selecting means 2 to combine the motion data of the model from the angle data of each joint (s15). ).

FIG. 18 shows the details of the processing procedure by the synthetic motion generating means. In the present embodiment, the combination of the characteristic motions such as "tired" and "cold" with respect to the basic motion is executed for each joint, and then the probabilistic motion is combined with the designated joint.

First, it is determined whether or not synthesis processing has been performed for all joints (s150). When the synthesis of all joints is not completed, the DC component of the FFT data is acquired from the motion database 21 (s151). The acquired FFT data is the FFT data (s1) of the current joint of the motion data selected as the optimum by the selection means 2 in response to the designation of the motion schedule.
51-1), FFT data of "tiredness" (s151-
2), FFT data (s151-3) of "cold". The obtained three-component FFT is parameter-synthesized (s
152). At this time, the “tired” FFT is multiplied by the “tired synthesis ratio” and the “cold” FFT is multiplied by the “cold synthesis ratio”, and then added.

Similarly, the AC component of the FFT data is acquired (s153-1 to s3-1), and parameter synthesis is performed (s1).
54). In this embodiment, the same composition ratio is used for DC and AC. Next, the synthesized DC FFT and AC FF
The inverse Fourier calculation is performed on the T data to restore the joint angle data (s155). The above process is repeated for all joints of the model.

After the angle data is restored for all joints, it is judged whether or not a stochastic motion is set for the selected motion (s156). If set, for the action (eg, “waving”) selected in the probabilistic action selection process (s12), the angle information of the joint designated for that action is obtained from the RAW data section 213 of the action database 21. The acquired joint information (s157) is replaced with the stochastic movement angle information from the angle data previously restored by the FFT inverse transformation for the designated joint and its child joints (s158).

By this processing, all the angle information of each joint of the designated model in the current frame including the probabilistic motion is determined. When the motion of each joint is determined, the angle information of all the joints is sent to the movement amount generating means 4.

The movement amount generating means 4 determines the movement amount and direction of the model from the movement path designated by the user and the combined movement of the model. FIG. 19 shows the details of the processing procedure of the movement amount generating means.

First, the synthesis result of the sent model and 1
The synthesis result before the frame is compared (s20), and how much the current synthesis result has moved from the one frame before is determined from the comparison result (s21). The movement amount is calculated from the difference between the positions of the joints with the grounded foot as a reference.

Next, it is judged whether or not the movement amount subtraction is designated in the synthesis operation selection process (s13) depending on the environmental conditions (s22). When the subtraction is designated, the calculated movement amount is multiplied by the subtraction ratio to perform the subtraction (s23).

Next, the coordinates of the target point on the operation path are acquired (s24), and one frame of the correction angle from the front direction of the current (processing target) model toward the target point is calculated (s).
25). Furthermore, the synthetic operation selection process (s
In 13), it is determined whether or not there is a shift designation for the motion path (s26), and if there is a shift designation, the shift amount (in the above, the correction angle is + 20%) is added (s27).

Finally, the position of the model in the virtual space of the current frame is corrected by adding the calculated movement amount to the position and angle information of the uppermost joint of the model in the previous frame and correcting the displacement angle with respect to the target point. And the direction is determined (s28). As a result, the position and orientation of the model are changed, and the model is moved and displayed on the screen of the display means 5.

For example, assuming that a curve is designated as an operation path, description will be made with reference to FIG. When moving a curved path, between the start point 201 and the end point 202, for example, virtual target points 203 to 201 corresponding to the stride of the model
Set 1. Model orientation 201 at start point 201
2 starts the correction toward the closest target point 203 after the movement amount is calculated, and the orientation 2013 is sequentially determined according to the number of frames expressing one step from the start point 201 to the virtual target point 203.
~ 2016 will be changed. Then, the virtual target point 203
When reaching the vicinity of, the virtual target point 204 is changed to the next virtual target point 204.

FIG. 21 shows a schematic diagram of the animation generated by this embodiment. In this example, as a schedule, "motion: walking", "frame number: 80", "motion path: path A", "stochastic motion: raising hands (frequency: 50)
%) ”,“ Model characteristics: male, 24 years old, physical fitness 5 ”, and“ environment setting: daytime, temperature 20 degrees ”.

The model 212 "walks on the path A 211.
Walking with FFT data of "male, 20 years old".
If there is no 20-year-old data in the database, nearby motion data is selected. Then, a random number is generated every two frames, and in a frame in which a probabilistic motion is determined, a “hand-raising” motion 213 is performed while walking on the path A. Also,
Therefore, the number of frames to be processed in the motion schedule increases, so that the combined ratio of “tiredness” increases and gradually changes to the “forward leaning” walking posture 214. Furthermore, a state transition occurs in the frame where the physical strength value becomes 0, and the basic motion is changed to “stop” 215 to stop the motion, and “tired”
The composition ratio of the sword increases and it becomes a "sag" posture.

As described above, in the present embodiment, even when exactly the same operation schedule, model characteristic, environment characteristic and stochastic operation are designated, the finally generated animation is different every time depending on the generated random number. , You can avoid uniform animation.

Even if the same schedule is designated, the generated animation can be variously changed by changing part of the model characteristic, the environmental characteristic or the stochastic operation. Moreover, changes can be easily made by simply selecting and designating on the menu screen of the user interface.

Further, since a general language is designated when selecting an operation, even a person who is not an expert in computer animation can create a complicated animation.

[0073]

According to the present invention, only by setting an operation schedule, the basic operation is modified or the movement amount is modified according to the model characteristic or the environmental characteristic, the stochastic operation is added, and the basic operation itself is transitioned. Since the animation played by the model can be changed in various ways, there is an effect that a natural animation can be created in a short time.

Further, since it is possible to autonomously determine the change or composition of the motion in each frame only by setting the motion schedule from the input screen of the user interface, there is an effect that the user can easily create and change the animation. is there.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an animation generation device (computer graphics) according to an embodiment of the present invention.

FIG. 2 is a data configuration diagram of an operation database according to an embodiment.

FIG. 3 is an explanatory diagram showing a parent-child relationship with a model joint structure.

FIG. 4 is an explanatory diagram showing a description example of cyclic operation data.

FIG. 5 is an explanatory diagram showing a description example of RAW operation data.

FIG. 6 is a screen configuration diagram of a user interface according to an embodiment.

FIG. 7 is a screen configuration diagram of an operation data selection menu according to an embodiment.

FIG. 8 is a screen configuration diagram of an operation path selection menu according to an embodiment.

FIG. 9 is an explanatory diagram of an operation path.

FIG. 10 is a screen configuration diagram of a stochastic operation selection menu according to an embodiment.

FIG. 11 is a screen configuration diagram of a model characteristic input menu according to an embodiment.

FIG. 12 is a screen configuration diagram of an environmental status setting menu according to an embodiment.

FIG. 13 is a schematic flowchart showing an operation selection process according to an embodiment of the present invention.

FIG. 14 is a detailed flowchart of operation data selection processing.

FIG. 15 is a detailed flowchart of probabilistic action selection processing.

FIG. 16 is a detailed flowchart of a combining operation selection process according to environmental conditions.

FIG. 17 is a detailed flowchart of a process of selecting a motion transition based on model characteristics.

FIG. 18 is a flow diagram of a combined motion generation process according to an embodiment of the present invention.

FIG. 19 is a flow chart of movement amount generation processing according to an embodiment of the present invention.

FIG. 20 is an explanatory diagram illustrating a movement amount generation process.

FIG. 21 is an explanatory diagram showing an example of a generated animation according to the present embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Database, 2 ... Operation selection means, 3 ... Composite operation generation means, 4 ... Movement amount generation means, 5 ... Display means, 6 ... Operation schedule table, 7 ... Stochastic operation table, 8 ...
Model characteristic table, 9 ... Environmental state table, 10 ... User interface, 21 ... Operation database, 22
... model characteristic database, 23 ... joint structure database, 24 ... motion transition database, 211 ... cyclic motion part, 212 ... feature amount data part, 213 ... RAW motion data part, 214 ... motion transition table, 215 ... motion category, 216 ... model characteristics, 217 ... joint structure,
218 ... Actual operation data, 61 ... Operation selection button, 62 ...
Selection operation display area, 63 ... Frame number designation display area, 6
4 ... operation path selection button, 65 ... operation path display area, 6
6 ... Probabilistic action selection button, 67 ... Probabilistic action identification number display area, 68 ... Model characteristic setting button, 69 ... Model characteristic identification number display area, 610 ... Environmental state setting button,
611 ... Environmental state identification number display area, 612 ... Schedule registration button, 613 ... Schedule change button, 6
14 ... Schedule delete button, 615 ... Schedule display area, 616 ... Total frame number display area, 617 ... Animation generation button, 618 ... End button, 101
... Probabilistic motion identification number input display area, 102 ... Stochastic motion selection area, 103 ... Frequency input display area, 104 ... Stochastic motion list display area, 105 ... Stochastic motion registration button, 1
06 ... Probabilistic action deletion button, 107 ... Probabilistic action setting menu end button, 111 ... Model characteristic identification number input display area, 112 ... Model characteristic input area, 113 ... Model type, 114 ... Model sex, 115 ... Model Age, 116 ... Physical strength of model, 117 ... Physical condition of model, 1
18 ... Model characteristic setting menu end button, 121 ... Environmental condition identification number input display area, 122 ... Environmental condition input area, 123 ... Temperature, 124 ... Luminous intensity, 125 ... Wind power, 12
6 ... End button of the environmental condition setting menu.

Continuation of the front page (56) Reference JP-A-11-120378 (JP, A) JP-A-9-311947 (JP, A) JP-A-4-195476 (JP, A) Muneto Unuma, 1 person, "Computer Method for Generating Human Walking Motion with Emotion in Animation ”, IEICE Transactions on Information and Systems II-Information Processing, The Institute of Electronics, Information and Communication Engineers, August 25, 1993, J76-D -Vol. II, No. 8, p. 1822-1831 Seiki Inoue, 1 person, "Image database for screen synthesis ~ Screen synthesis by moving image parts ~", Proc. Of the 39th National Conference (Late 1989) (II) (Data Processing Software Software Engineering), Japan Society for Human Information Processing, October 18, 1989, p. 964-965 (58) Fields investigated (Int.Cl. ⁷ , DB name) G06T 15/70 CSDB (Japan Patent Office)

Claims (11)

(57) [Claims] 1. A periodic motion data for a plurality of models.
Data and characteristic motion data that follows its cyclic motion.
From the accumulated motion database, desired motion and route
And animate consisting of multiple frames by specifying the model
In the method of generating a model, a model consisting of a motion name, model specific characteristics and capability characteristics.
The operation schedule that specifies the video characteristics and the number of frames.
Unique with the specified operation name from the above operation database
Optimal periodic motion data and reflection of specified ability characteristics
Select the first characteristic motion data adapted to
The operation data is reflected in all frames and the latter operation
The data can be changed by evaluating the change in the performance characteristics for each frame.
However, when combining the former and latter motion data for each frame,
In this case, the specification of the operation schedule includes the environmental characteristics in which the model exists, and the second characteristic operation data that is adapted to the specified environmental characteristics is selected from the operation database, and the characteristic operation data is selected for each frame. A method for generating an animation, characterized in that the change in the environmental characteristic is evaluated and varied, and the change is combined with the other motion data described above for each frame. 2. The model according to claim 1, wherein the intrinsic characteristic of the model includes type (eg, human), gender, etc., the performance characteristic includes physical strength, etc., and the environmental characteristic includes temperature, etc. "Combined ratio of fatigue" due to the decrease in physical strength
Is calculated, and the first characteristic motion data representing "fatigue" is changed according to the combined ratio, and / or the "cold combined ratio" due to the decrease in the temperature is calculated, and according to the combined ratio. A method for generating an animation, wherein the second characteristic motion data representing "cold" is changed. 3. A plurality of frames by designating desired motions, paths, and models from a motion database accumulating periodic motion data of a plurality of models and stochastic motion data following the periodic motions. In the method of generating an animation consisting of, a motion schedule that specifies a motion name, model characteristics, a probabilistic motion name and frequency, and the number of frames is referred to from the motion database, and the optimum cyclic motion for the specified motion name and model characteristics. Data is selected and reflected in all frames, and whether or not a probabilistic operation is performed is determined from the specified frequency for each frame, and if it is performed, the probabilistic operation name specified in the periodic operation data of the frame is used. A method for generating an animation characterized by synthesizing motion data. 4. The periodic motion data according to claim 3 , wherein the periodic motion data includes “run”, “walk”, “stop”, etc., and the stochastic motion data includes “raise hand”, “shake head”, etc. , When one action name of the former is set to a plurality of action names of the latter in the action schedule, one of the probabilistic action names and the necessity of execution thereof are determined by the generation of random numbers. How to generate an animation. 5. A method for generating an animation composed of a plurality of frames by designating a desired motion, a path, and a model from a motion database accumulating periodic motion data for a plurality of models for each motion name, In the operation database, a plurality of operation transition sequences indicating the order of operation transitions are stored in advance, and an operation name, an operation schedule in which a model characteristic consisting of a characteristic characteristic and a capability characteristic of the model and a frame number are designated are referred to, From the motion database, select the motion transition sequence that starts with the specified motion name along with the specified motion name and the cyclic motion data that is most suitable for the unique characteristic, and create a frame starting from the selected cyclic motion data. When the change in the above-mentioned performance characteristics is evaluated for each frame and the change exceeds the threshold value, the next action name of the selected action transition sequence A method for generating animation, characterized by changing to periodic motion data to create a frame. 6. The method according to claim 5 , wherein an environmental characteristic is specified in the operation schedule, and when the capability characteristic of the model and / or the environmental characteristic changes every frame and exceeds a threshold value, the periodic characteristic is set. A method for generating an animation characterized by changing motion data. 7. A method of generating an animation composed of a plurality of frames by designating a desired motion, a path and a model from a motion database accumulating a plurality of periodic motion data of a plurality of models, wherein: The behavioral database stores characteristic behavioral data that follows the periodical behavior, probabilistic behavioral data, and behavioral transition sequences that indicate the transitional order of the periodical behavior, and is a model consisting of behavioral names, unique characteristics, and ability characteristics. Characteristic,
By referring to an operation schedule specifying a probabilistic operation including environmental characteristics and frequency, periodical operation data that is optimum for a specified operation name and unique characteristics is selected from the operation database, and the performance characteristics and // Alternatively, the characteristic motion data adapted to the environmental characteristics are changed according to changes in those characteristics, and it is determined whether or not the probabilistic motion is required to be added based on the designated frequency, and it is determined that the changed characteristic operation data or the addition is required. The selected stochastic motion data is combined with the selected cyclic motion data, and further, when the change of the performance characteristic and / or the environmental characteristic exceeds the threshold value set for each, the motion transition sequence A method of generating animation, characterized in that the periodic motion data is changed in order. 8. The method according to claim 7 , wherein between the previous frame and the current frame, the cyclic motion data or the composite motion data in which the cyclic motion data and the other motion data are combined is compared, and drawing is performed from the difference. The method for generating an animation is characterized in that the movement amount is corrected by a subtraction value of the movement amount in accordance with the change in the environmental characteristic when the movement amount for is calculated. 9. A database accumulating a plurality of motion data for a plurality of models, a user interface for designating a motion schedule including a motion, a path, a model, the number of frames, etc. for an animation to be created, and a designated motion. It has a movement selection means for selecting movement data to be used based on a schedule and a movement amount generation means for obtaining a movement amount of movement data between frames,
An animation generating device for displaying a frame screen according to the amount of movement, and a computer graphic .
In the animation creating device , the database includes at least a plurality of periodic motion data of a basic motion as the motion data, and characteristic motion data that follows the cyclic motion or stochastic motion data that is stochastically added. The user interface stores one, the model characteristic table for setting the intrinsic characteristic and the capability characteristic of the model for the schedule, the stochastic operation table for specifying the probabilistic operation and its frequency, or the model At least one environment state table for setting environment characteristics of the existing virtual space is provided, and the operation selecting unit selects optimum cyclic operation data based on the operation name and unique characteristics set in the operation schedule. And determining the addition of stochastic motion based on the frequency or the performance characteristic or the Perform at least one variable of the characteristic operating data based on the change in the boundary characteristics, animation using computer graph acme box, characterized in that a resulting function of combining the optimal periodic motion data of the data and the by Creation device
Place 10. The database according to claim 9 , wherein the database stores a motion transition sequence indicating a change order in the case of transitioning a periodic motion, and the motion selection means sets a change in the capability characteristic to a predetermined value. An animation by computer graphics, which is provided with a function of transitioning to the next periodic operation of the action transition sequence starting with the selected periodic action when exceeding the limit.
Creation device . 11. The animation creation by computer graphics according to claim 9 or 10 , wherein the user interface is configured to have an input screen of the operation schedule and a menu screen corresponding to setting contents. Equipment .