JPH08123976A - Animation generating device - Google Patents

Abstract

PURPOSE: To save the trouble to edit an animation by acquiring information, required to edit the animation, from a scenario wherein the contents of the animation are described automatically to some extent. CONSTITUTION: This device is equipped with a scenario analysis part 113 which analyzes the structure of scenario sentences by using the knowledge of a scenario analytic dictionary 112, an animation information extraction part 115 which extracts animation information from the analytic result and stores it in an animation information storage part 114, an animation editing part 120 which performs the necessary editing of the animation information stored in the animation information storage part 114 and the addition of new information, and an animation control part 121 which sends sound information to a voice synthesis part 109 and image information to an image display part 108 while controlling a voice signal processing part 110 on the basis of the stored animation information.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for creating animation using a computer, and in particular, a part of an animation creating work based on the scenario information, using information written in the word scenario as input. The present invention relates to an animation creating device capable of efficiently creating an animation by automating.

[0002]

2. Description of the Related Art Animation has long been popular with people as one of applications that make full use of multimedia. However, it requires a special ability to draw a slightly different picture for each frame, and it requires enormous effort to create a large number of cel images (about 16 frames per second). Therefore, the problem was that no one could easily make it.

However, in recent years, due to the dramatic improvement in the processing capability of computers, it has become possible for computers to handle not only character information but also various information media such as voices, still images, and moving images. With this,
Various methods of creating animation using a computer have been developed, and a representative one is an animation creating apparatus based on a film model in which a cel image is a unit.

An animation creating apparatus based on a film model realizes on a computer an environment for creating an animation, which is performed on a conventional desk, in which a slightly different picture is drawn for each frame of a film forming an animation. That is, first, a scenario is written, the content of the animation is thought in mind while referring to the scenario, and the frame is drawn frame by frame in a cell displayed on the computer screen.

This method is very simple, and it is possible to immediately reproduce the created cel image as an animation and check its contents. It is easy to correct the cel image by using a drawing tool, especially for consecutive frames. When inserting a new frame in between, by making full use of image processing technology, it is possible to semi-automatically generate a picture of the frame to be inserted, and it is possible to make animation creation more efficient by utilizing the power of the computer. it can. Like this,
In the conventionally developed animation creation system based on the film model, the efficiency of drawing each cell that composes the animation is improved as compared with the conventional handwriting.

However, the procedure for creating an animation is basically the same as in the case of hand-drawing, in which each frame is drawn based on a scenario. That is,
From the scenario first described in the language, it is necessary to manually draw the moving images and background images of the characters required for animation, and edit the order in which those images appear in the animation. In some cases, images used as materials for animation are stored in the database, but even in that case, the conventional animation creation system uses the file name corresponding to each data when reusing the images. Only the method of obtaining the original data is provided, and as the number of data files increases, there is a problem that the time and effort required to search for desired data also increase. That is, it is not taken into consideration that a part of the animation created using the conventional system can be reused in the creation of another animation and the efficiency of the animation creation is further improved.

An animation creating apparatus based on an object model is a method developed from the viewpoint completely different from the above film model. This method attempts to automatically reproduce an animation by giving knowledge to each of the appearances in the animation and inferring the behavior of the appearance using this knowledge. Specifically, the behavior of each appearance is defined by some parameters that determine it. The animation creator sets appropriate initial values for these parameters for each appearance. After the parameters are set, each appearance recognizes the surrounding situation, infers the behavior according to the situation and the value of the parameter set for itself, and automatically generates the behavior. As described above, this method has an advantage that an animation can be generated only by setting parameters that determine the behavior of each appearance thing.

However, when an animation is created by this method, there is a problem in what value the parameter is set to. That is, the above method is suitable for creating an animation in which an initial object is given a certain initial value, and then the appearance object arbitrarily decides the surrounding situation, but the animation scenario is already used. Has been decided, and when creating an animation based on it, it is difficult to set the parameters of each of the appearance items so as to match the contents of the scenario, so create an animation based on the scenario. In some cases, it is not appropriate to apply this method.

As a conventional development example of the animation creating system based on the film model, "MacriMind Director" of Macromind (MacriMind) Co., Ltd., and as an animation creating method based on the object model, For example, refer to [Sakamoto, Hamada, John Z., "Interactive animation description language and its processing system" AV-Script "",
IPSJ Research Report, 92-IM-6, Vol.92, No.34, pp.2
5-32, May 1992].

[0010]

When an animation is created according to the contents of a scenario, in order to further improve the efficiency of the creation work, a part of the work manually performed in the conventional system is automated. That is a challenge. Specifically, instead of manually drawing each cell that is the source of the animation from the scenario described in the language, the scenario described in the natural language is analyzed using natural language processing technology, Referring to the analysis result, search for a still image used as a background of animation and a moving image used as a character, the order in which the searched image appears in the animation,
By automating the lines of the characters, it is an issue to improve the efficiency of the animation creation work.

In addition, since an animation is usually created by superimposing an image showing the movement of a character on a background image, a background image as a stage background, a foreground image used as a stationary prop for the stage, and movement are involved. It is effective to convert the character image into data, save it in a disk as a file, and reuse the file. Therefore, it is an object to make the storage space efficient when storing the file in the database and to facilitate the search of the file stored in the database.

The present invention solves the above-mentioned problems, and reduces the trouble of editing animation by automatically obtaining the information necessary for editing the animation to some extent from the scenario describing the contents of the animation. It is an object of the present invention to provide an animation creating device that does.

[0013]

In order to achieve the above object, the present invention uses a language dictionary for storing information necessary for analyzing a natural language and a natural language using the information stored in the language dictionary. A natural language processing unit that analyzes the image, an image database that stores the images used for animation,
An image database management unit that registers an image used for animation in the image database and searches from the image database, an image display unit that displays the image data, and inputs a natural language to convert the natural language into a voice signal. A voice synthesis unit, a voice signal processing unit that processes a voice signal, a scenario input unit that inputs a scenario sentence, a scenario analysis dictionary that stores knowledge necessary for analyzing the structure of the scenario sentence, and the scenario analysis unit. From the scenario analysis unit that analyzes the structure of the scenario sentence using the knowledge stored in the dictionary, the animation information storage unit that stores the animation information that is the information necessary for playing the animation, and the analysis result of the structure of the scenario sentence. Animation for extracting the animation information and storing it in the animation information storage unit The information extraction unit and the animation information stored in the animation information storage unit and the animation editing unit that edits the animation information stored in the animation information storage unit if necessary and adds new information. While controlling the audio signal processing unit based on the above, as an information necessary for reproducing the animation, an audio control unit for sending audio information to the audio synthesis unit and image information to the image display unit, respectively.

[0014]

Operation: The creator of the animation inputs the scenario in which the contents of the animation to be created are written in the scenario input section. The input scenario is sent to the scenario analysis section, and the structure of the scenario is analyzed. Subsequently, the animation information extraction unit extracts animation information necessary for reproducing the animation from the analyzed scenario structure. Further, the creator of the animation adds the information that is insufficient for the reproduction of the animation while referring to the extracted animation information, and changes the information to complete the animation information.

Finally, the animation information is sent to the animation control unit, and the animation control unit controls the audio signal processing unit on the basis of the animation information in the animation information storage unit to obtain the information necessary for reproducing the animation. , Voice information to the voice synthesizer and image information to the image display unit. The voice synthesis unit converts the input character string into a voice signal, the voice signal processing unit converts the voice signal, and the voice output unit outputs the converted voice. In the image display unit, the background image and the foreground image are combined and output according to the reproduction speed.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT An embodiment of an animation creating apparatus according to the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of an animation creating apparatus according to an embodiment of the present invention.

The animation creating apparatus of this embodiment comprises a language dictionary 101, a natural language processing section 102, and an image data storage section 10.
3, anime brush data storage unit 104, image search table 10
5, an image database 106, an image database management unit 107, an image display unit 108, a voice synthesis unit 109, a voice signal processing unit 110, a scenario input unit 111, and a scenario analysis dictionary 11
2. Scenario analysis unit 113 and animation information storage unit 11
4, animation information extraction unit 115, animation editing unit 120, animation control unit 121, image input unit 122, feature extraction unit 123, image cutout editing unit 124, audio output unit 125
Consists of

As shown in FIG. 13, the animation editing unit 120 is composed of a target moving route editing unit 116, a target moving distance calculating unit 117, a target moving speed editing unit 118, and a target moving speed calculating unit 119. .

Next, the contents of the processing in each of the above-mentioned parts constituting the animation creating apparatus will be described. First, the animation creating device image database 106 includes an image data storage unit 103, an animation brush data storage unit 104, and an image search table 105. FIG. 2 shows an example of the configuration of the image database 106. FIGS. 2A, 2B, 2D, and 2E show the image data storage unit 1 in this order.
A conceptual diagram showing the contents of the image data stored in 03,
4A and 4B are diagrams showing a description example of an animation brush data storage unit 104 and a description example of an image search table 105.

The above-mentioned three components of the image database 106 will be described below. The image data is data that includes pixel data that describes information about each pixel and attribute information that describes a target attribute and a file name as information about image content. The image data storage unit 103 stores the image data. .

For example, in the example of FIG. 2A, three image data are shown, but the target attributes are all "Parts",
Also, the file names are from the left in order `` Kenchan-1.img, Kench
an-2.img, Kenchan-3.img ”. The above "P
“Arts” indicates that the image data represents a foreground target, and if the image data represents a background target, it can be distinguished from the foreground target by describing “BackGround” as a target attribute. You can do this. The meanings of the foreground object and the background object will be described later in the explanation of the animation information extracting unit 115. Since the method of recording the attribute information of the image data and the pixel data is a known technique, the description of the recording method will be omitted.

Here, the animation brush data is data in which a set of image data forming an animation is ordered and put together, or a set of pointers to image data forming an animation is set in order. The data collected as one is stored in the anime brush data storage unit 104.

For example, in the example of FIG. 2B, the first column in the figure is the name of the animation brush data, and the second column is a pointer to a series of image data constituting the animation brush data. Is. This is the image data Kenchan-
1.img, Kenchan-2.img, Kenchan-3.img, Kenchan-2.img
By displaying in this order, it is possible to animate the subject "Ken no Ken-chan runs". FIG. 2C shows a conceptual diagram depicting the contents of the animation brush.

In the following, both the above-mentioned anime brush data and image data will be collectively referred to as an image.

The image search table 105 is composed of description parts of a search keyword, an animation brush data name, and an image data name. For example, in the first line of the example shown in FIG. 2D, "Kenchan" and "Kenchan-ru" are displayed in order from the left column.
n-1.mv, ... '' `` Kenchan-1.img, Kenchan-2.img, Kencha
n-3.img, Kenchan-4.img, ... ". This is the image data to which the search keyword "Ken-chan" is added.
run-1.mv etc. Also, in the image data, Kenchan-1.im
g, Kenchan-2.img, Kenchan-3.img, Kenchan-4.img, etc. are meant.

Further, in the image search table 105, based on the rule that "the image data that constitutes the animation brush data described in the animation brush data description section is not described in the image data description section", FIG. It is possible to reduce the amount of data to be described in the image search table 105 shown in (d).

FIG. 2 (e) shows an example of the image search table 105 in which the description amount is reduced. 2D and 2E are examples of the image search table 105, the difference between them is the description content of the image data storage unit. That is, comparing the image data storage units in the first row of both figures, in the case of FIG.
nchan-1.img, Kenchan-2.img, Kenchan-3.img, Kenchan
-4.img, ... ", in the case of Fig. 2 (e)" Kenchan-4.img, ... "
2c is compared with FIG. 2d, and Kenc is shown in FIG. 2e.
The three image data of han-1.img, Kenchan-2.img, Kenchan-3.img are omitted. This is the anime brush data Kenchan-run-1.mv of the above 3 image data in Fig. 2 (e).
This is because it can be obtained by referring to the image data description part of.

The image database management unit 107 searches for an image necessary for creating an animation in response to an image search request from the animation editing unit 120, and sends the search result to the animation editing unit 120. Further, when a request for data registration is made by the image cutout editing unit 124, the data registration process is performed.

The flow of image retrieval and registration processing will be described below with reference to FIGS. 3 and 4, respectively.

<1. Image Retrieval> FIG. 3 is a flow chart showing the flow of processing for retrieving an image in the image database management unit 107.

(301) Input an inquiry sentence in which the content of the image to be searched is described in natural language, the target attribute of the image (ie, foreground target or background target), and the operation attribute of the image to the image database management unit 107. To do. However, the action attribute is the image database management unit that indicates which animation brush data or image data is desired as a search result.
This is an attribute value for notifying 107.

(302) The input inquiry sentence is sent to the natural language processing unit 102, the input sentence is morphologically analyzed, and the input sentence is divided into morphemes.

(303) Of the divided morphemes, only the morphemes to be the image search keyword are extracted. Whether a certain morpheme becomes an image search keyword can be determined by checking whether the morpheme is registered in the image search table 105 as a search keyword.

(304) An image related to the keyword is searched. Regarding the image search method using keywords,
For example, reference [Shibata, Inoue, "Associative search method for image database 106", IEICE Transactions, Vol.J73-D-
II, No. 4, pp. 526-534, April 1990].

(305, 306) As a result of the search, if there is no image representing the contents of the input inquiry sentence in the image data storage unit 103, the process is terminated. When there is a search image, only the image that matches the input target attribute and motion attribute is selected from the searched images, and the selected image is set as the result of the main image search process.

<2. Image Registration> FIG. 4 is a flow chart showing the flow of processing for registering an image in the image database management unit 107.

(401) A registered image, a sentence describing the contents of the registered image in natural language, and a target attribute and a motion attribute of the image are input to the image database management unit 107.

(402) A sentence describing the contents of the image in natural language is sent to the natural language processing unit 102, and morphological analysis of the sentence is performed to divide the sentence into morphemes.

(403) From the divided morphemes, only the morphemes that are image search keywords are extracted. As a method for checking whether or not a certain morpheme serves as an image search keyword, the method described in the flow of the image search processing may be used.

(404) It is checked whether the input operation attribute is the animation brush data. Then, if the input operation attribute is the animation brush data, the following (a) animation brush data registration processing is performed. If the input operation attribute is not the animation brush data, the following (b) image data registration processing is performed.

(A) Animation brush data registration processing (405) Regarding a series of image data constituting an animation brush, image data not registered in the image data storage unit 103 is registered as image data whose target attribute is the foreground target. .

(406) The animation brush data is registered in the animation brush data storage unit 104. (407) For each keyword extracted in the process of 403, the animation brush data is additionally registered in the description part of the animation brush data of the image search data of the image search table 105 having the same search keyword as the keyword. Then, the animation brush data registration process ends.

(B) Image data registration processing (408) It is checked whether the input target attribute is the foreground target.

(409) If the input target attribute is the foreground target,
The input image is registered in the image data storage unit 103 as a foreground target.

(410) If the input target attribute is the background target,
The input image is registered in the image data storage unit 103 as a background target.

(411) For each keyword extracted in the process of 403, the image data is added to the image data description part of the image search data of the image search table 105 having the same search keyword as the keyword. Register and finish the image data registration process.

When the registration process of the animation brush data or the image data is completed as described above, the image database management unit
The image registration process at 107 is terminated.

The image input unit 122 reads a printed matter (for example, a cartoon) on which image data as a material forming each scene of animation is drawn, and performs preprocessing for converting the data into computer usable data. so,
For example, a scanner device or the like is used.

The image cutout editing unit 124 performs a process of cutting out image data as a material forming each scene of animation from the image input by the image image input unit 122. Also, make a copy of the cut out image data,
The work of creating the animation brush data is also performed by correcting the duplicated image data.

The feature extraction unit 123 is an image cutout editing unit 124.
For each of the image data and the animation brush data cut out in step 2, if it is necessary to extract the features of the two types of data when registering them in the image database 106, for example, compressing the two types of data as they are, The feature extraction process is executed.

The natural language processing unit 102 uses the language dictionary 101 to analyze natural language sentences. Particularly, in the present embodiment, as the analysis result of the input natural language sentence, the noun which is the main body of the input natural language sentence and the time required for uttering the input natural language sentence are required in the animation information extraction unit 115. This embodiment is characterized by extracting one piece of information. In this embodiment, a method of extracting the above two pieces of information will be described.

<1. Extraction of Main Noun in Input Natural Language Sentence> In the present embodiment, the main noun in the input natural language sentence means a noun applicable to the following two cases.・ The head of the noun phrase when the input natural language sentence is a noun phrase ・ The action subject of the sentence when the input natural language sentence is a sentence is the root of the meaning of the noun phrase in the noun phrase An eggplant noun. For example, for the noun phrase "mouse no Ken-chan", the noun phrase "ken-chan" is a modifier to characterize the next "ken-chan", while the noun phrase "ken-chan" means And the noun "ken-chan" is the head of the noun phrase.

The action subject is a noun that is the subject of the action in the sentence. For example, for a sentence "mouse no Ken-chan is dancing", "mouse no Ken-chan ga" becomes the main part of the sentence, and the head of the noun phrase "mouse no Ken-chan" that represents the subject of the main part. "Ken-chan"
Is the subject of action.

FIG. 5 is a flow chart showing the flow of processing for extracting nouns corresponding to the two said head and action subjects.

(501, 502, 503) The sentence input to the natural language processing unit 102 is first decomposed into morphemes, which are the smallest units constituting the sentence, by morphological analysis, and then by syntactic analysis,
Parse the syntactic structure of the input string. Regarding the technology of the above morphological analysis / syntactic analysis and the construction of the language dictionary used in the analysis, see [Hirogo Nomura, "Basic Technology of Natural Language Processing", The Institute of Electronics, Information and Communication Engineers, 1988], etc. Since it is disclosed as a publicly known technique in the field of natural language processing, description thereof will be omitted.

(504,506) By comparing the syntactic structure of the analysis result of the structure of the input sentence with the grammatical rules of the language dictionary 101, the syntactic structure is a sentence consisting of a noun phrase, a main part and a predicate, or other than the above. Determine which one.

(505) When the input sentence is a noun phrase, when the sentence to be input is decomposed into morphemes, the noun located at the rightmost end is extracted as the head of the noun phrase.

(507,508) When the input sentence is a sentence composed of a main part and a predicate, the main part is first extracted. Subsequently, the main part is divided into a noun phrase and a case particle connected to the noun phrase,
The head of the noun phrase is extracted as the subject of the sentence.

(509) When the input sentence is neither a noun phrase nor a sentence consisting of a main part and a predicate, the main language noun cannot be extracted from the input sentence in the module that calls the natural language processing unit 102. Let me know. With the above, the natural language processing unit 102
The process of extracting the noun that is the main subject in the input natural language sentence is completed.

In addition, in (505) of the process of extracting the subject from the above noun phrase, "When the noun phrase is decomposed into morphemes,
The noun located on the far right is the head of the noun phrase. ”However, in order for this process to be executed correctly, only the phrase in which the last word is composed of a noun should be accepted as a noun phrase. , It should be described in the language dictionary 101. That is, by rewriting the grammatical rules for parsing described in the language dictionary 101, it is possible to meet the requirement of what kind of noun is regarded as the head. In this example, the rule is applied to Japanese.For example, in the case of English, not only the rewriting of the grammar rule but also the content of (505) of the process of extracting the head from the above noun phrase is adjusted to English grammar. Need to change.

<2. Time required for utterance of input natural language sentence> The time required for utterance of the input natural language sentence represents the time required from the beginning of utterance of the input natural language sentence to the end thereof. FIG. 6 is a flowchart showing the flow of processing for extracting the time required for uttering an input natural language sentence.

(601, 602) Similar to the processing of <1.>, the sentence input to the natural language processing unit 102 is first decomposed into morphemes, which are the minimum units constituting the input sentence, by morphological analysis.

(603) Total time T required for utterance of input sentence T
Set s to 0. (604) For each of the decomposed morphemes, the time required for uttering the morphemes is obtained by examining the language dictionary 101 and added to the Ts.

(605) Ts is notified to the module that called this process as the time required to utter the input sentence, and the process of calculating the time required to utter the input natural language sentence in the natural language processing unit 102 is ended.

In the above 604, as a method of constructing the natural language processing unit 102 and the language dictionary 101 for extracting the time required for uttering a morpheme, for example, when handling Japanese, the word As a set of headings representing the pronunciation of the word, the correspondence between the morpheme heading and the kana reading heading of the morpheme is described as shown in FIG. 7 (a), and the word pronunciation according to the present invention is represented. As a means for calculating the time required to pronounce the word from the headline, a method of setting the length of the character string of the kana reading headline as the time required to utter the morpheme may be used. Further, when dealing with English, in the language dictionary 101, as a set of a word heading and a time required to pronounce the word,
As in (b), for example, the correspondence between the word heading and the time required for uttering the word may be described.

The scenario input section 111 describes the content of a desired animation in natural language. In the following, a sentence that describes the content of animation is called a scenario. FIG. 8 is a diagram showing an example in which the scenario input to the scenario input unit 111 is described in Japanese. Hereinafter, this scenario is referred to as scenario (1). The description content of the scenario (1) will be described in detail in the description of the scenario analysis dictionary 112 and the scenario analysis unit 113.

The scenario analysis dictionary 112 includes a scenario input section.
It describes the knowledge for analyzing the scenario edited in 111. FIG. 9 is a diagram showing an example in which the knowledge is defined based on the BNF notation. In FIG. 9, a character string consisting of _ (underscore) and English characters is called a sign, in particular, signs enclosed by $ <> $ are called nonterminals, and signs other than nonterminals are called terminal symbols. The meaning of each knowledge is that "the sign on the left side of" :: = "is the line of signs on the right side of" :: = "arranged in that order".

For example, in the case of knowledge G13, <speech>
Is the part in which <sentence>, [, <sentence>,] are arranged in this order. ”In the case of knowledge G14, <speech> is NARRATION_SPEECH, [, <sentence>,
] Are the parts arranged in this order. ” It should be noted that <sentence> in knowledge G13 is an arbitrary Japanese character string, and NARRAT in knowledge G13
ION_SPEECH represents the terminal symbol that represents the narrator's serif.

The scenario analysis unit 113 analyzes the structure of the scenario edited by the scenario input unit 111 using the scenario analysis dictionary 112. Using the scenario analysis dictionary 112 of FIG. 9, for example, knowledge G13 is applied to the analysis of the sentence on the fourth line of the scenario (1), the mouse Ken-chan [It is not necessary for the mouse to be a cat. Is <speech
The part corresponding to the non-terminal symbol> is parsed. Subsequently, the knowledge G8 is applied, and it is analyzed that the non-terminal symbol <speech> corresponds to the non-terminal symbol <scene_body>.

As described above, the analysis result of the scenario (1) analyzed using the scenario analysis dictionary 112 is sent to the animation information extraction unit 115 described below. The animation information storage unit 114 stores background information and foreground information as information for reproducing and displaying the animation.

In the present embodiment, the background object is used as the background information, and the foreground information is used as the foreground object, the foreground main object representing the subject of the foreground object, the foreground object speech, the speech time of the speech, and the foreground object. Eight sets of information of the appearance time, the moving path of the foreground target, the moving speed of the foreground target, and the voice quality of the foreground target are set as information necessary for reproducing the animation (hereinafter, individual information included in background information and foreground information). Is called animation information).

The meaning of each piece of animation information will be briefly described below. -Background target In the present embodiment, a target used as a stage background of a scene in animation, like "edge" in the example of scenario (1), is called a background target.・ Foreground target As in the example of scenario (1), “Ken no Ken-chan” and “Ken-chan Kentai”, they exist in the background of the scene in the animation and are accompanied by movement depending on the content of the animation story. In the present embodiment, a target that utters is called a foreground target. -Foreground main target The subject or subject of the foreground target is called the foreground main target. In the example of scenario (1), “Ken-chan” is the main target of the foreground for “Men's Ken-chan” and “Shemale Ken-chan”.・ Foreground target dialogue, speech time of foreground subject, appearance time of foreground subject of the contents corresponding to the foreground subject, the contents of the speech spoken during the appearance, the time to speak the dialogue, and the time to appear in the animation It means that.

In the following, when the word "object" is used, it means both the background object and the foreground object.

The animation information extraction unit 115 extracts animation information from the result analyzed by the scenario analysis unit 113, and stores the extracted information in the animation information storage unit 114.

Extraction of animation information is carried out in two cases: using the correspondence between the markers defined in the scenario analysis dictionary 112 and the animation information, and utilizing the animation information extracted using the correspondence. Do either method. Each method will be described below.

<1. When the Correspondence Relationship between Signs and Animation Information Defined in the Scenario Analysis Dictionary 112 is Used> In this method, the correspondence relationship between signs and animation information defined in the scenario analysis dictionary 112 is calculated. It is determined in advance, and animation information is extracted using the correspondence relationship. Hereinafter, first, the correspondence will be described, and then a method for extracting animation information using the correspondence will be described.

FIG. 10 is a diagram showing an example of a correspondence relationship between terminal symbols and non-terminal symbols in the knowledge for analyzing a scenario stored in the scenario analysis dictionary 112 and animation information. In this example, knowledge G1
In 3, the first sign <sentence> on the right side of “:: =” is “foreground target” which is one of the animation information, and the third sign <sentence> is “serif of the foreground target” which is one of the animation information. It has a corresponding relationship with. That is, when a sentence suitable for the knowledge G11 is extracted from the scenario during the scenario analysis, "::
= ”On the right side of the paragraph, the character string in the sentence corresponding to the first marker <sentence> is the character string in the sentence corresponding to the third marker <sentence> as the animation information foreground object. , Animation information It will be extracted as a serif for the foreground target.

For example, the knowledge G13 is applied to the analysis of the sentence on the 4th line of the scenario (1), the mouse Ken-chan [I don't need a cat to work on mice], and then, based on the above correspondence, The character string "rat no Ken" in the above sentence corresponding to the first sign <sentence> is the third sign <senten> as the animation information foreground target.
The character string in the above sentence corresponding to ce>"It is not necessary for the cat to work for rats" is extracted as the serif for the animation information foreground.

In this embodiment, as described above, by utilizing the correspondence between the markers defined in the scenario analysis dictionary 112 and the animation information, the lines of the background object, the foreground object, the foreground object of the animation information, Is extracted.

<2. Using Animation Information Extracted Using the Correspondence Relationship> Animation information that is not extracted using the correspondence relationship is derived from the animation information extracted using the correspondence relationship. The method of extracting each animation information is required individually.

A method of extracting the foreground main object, the speech time of the foreground object dialogue, and the foreground object display time among the animation information in this embodiment will be described below. -Speech production time of the foreground main subject and the foreground subject dialogue In the present embodiment, the time required for the speech of the foreground main subject and the foreground subject dialogue is the foreground subject and the foreground subject extracted in the process of <1.>. It is obtained by inputting the lines of the above into the natural language processing unit 102. However, foreground target without serif,
That is, for the foreground object that appears in the animation but does not speak the dialogue at all, the default value of the time required for the speech of the dialogue is determined in advance, and the value is used.

The calculation of the speech time of the foreground target speech according to the present embodiment is performed as follows. Let the foreground target for which the speech time of the speech is to be calculated be foreground target 1. The speech end time of the foreground object 0 appearing immediately before the foreground object 1 is set as the speech start time of the foreground object 1, and the time obtained by adding the time required for the speech of the foreground object 1 to the foreground object 1 to the foreground object It is decided to be the ending time of the speech of subject 1.

By performing the above calculation for the foreground objects in the order in which the foreground objects appear in the scenario, the speech utterance time for all the foreground objects appearing in the animation can be calculated.

By this method, for example, in scenario (1), the lines "Ken no Ken-chan""Idon't need a cat to work on rats" and "I am dressed as Ken-chan""Ora's mother bra no need- Assuming that the time required for uttering "" is 14 and 15, respectively, the speech time for "Rice no Ken-chan" can be calculated as 0-14, and the speech time for "Shemale Ken-chan" can be calculated as 14-29. Extraction of Display Time of Foreground Target For the display time of the foreground target, the speech time of the extracted serif of the foreground target is directly copied and used. That is, the display time of the foreground object and the speech time of the dialogue are exactly the same at the end of the process in the animation information extraction unit 115.

The explanation of each piece of animation information and the method of extracting the animation information have been described above.

Next, the flow of processing in the animation information extraction unit 115 will be briefly described with reference to FIGS. 11 (a) and 11 (b). FIG. 11A is a flowchart showing a rough flow of the entire process in the animation information extraction unit 115.

(1101) The scenario analysis unit 113 checks whether the background target has been extracted. (1102) If extracted, the background target is registered in the background information of the animation information storage unit 114.

(1103) The scenario analysis unit 113 checks whether a foreground target and a set of serifs for the foreground target have been extracted.

(1104) If it has been extracted, the following foreground information registration subroutine is executed for all the extracted pairs by inputting the foreground target and the serif of the foreground target in the order of extraction. To do.

The foreground object registration subroutine is for performing a process of derivatively extracting animation information from the lines of the foreground object and the foreground object. FIG. 11 (b) shows the flow of the process in the form of a flowchart. is there. The flow of processing of the subroutine will be briefly described below with reference to FIG.

(1105) The foreground target and the lines of the foreground target are registered in the foreground information. (1106) The structure of the sentence representing the input foreground target is analyzed. This analysis is processed by the natural language processing unit 102 described above.

(1107) As a result of the analysis, it is determined whether or not a main noun has been extracted from the foreground object.

(1108) When the main noun cannot be extracted from the foreground object, the foreground object is registered as the foreground main object in the foreground information.

(1109) When the main noun can be extracted from the foreground object, the noun is registered in the foreground information as the foreground main object of the foreground object.

(1110) It is checked whether or not there is a serif in the foreground object. (1111) If there is a serif in the foreground target, the serif is registered in the foreground information.

(1112) The natural language processing unit 102 calculates the time required for the speech of the speech, and then calculates the speech time of the speech with reference to the speech time of the speech of the foreground target extracted immediately before. Then, it is registered in the foreground information.

(1113) The dialog speech time of the foreground target is registered in the foreground information as the appearance time of the foreground target.

When the above processing is completed for all the extracted foreground objects and the serifs of the foreground objects, the processing in this subroutine is completed, and at the same time, the main routine of the animation information extraction processing is also completed.

The animation editing unit 120 manually performs the following two processes. -Correction of information extracted by the animation information extraction unit 115-Replenishment of information that cannot be extracted by the animation information extraction unit 115 As described above, the information extracted by the animation information extraction unit 115 includes background objects and foreground objects. , The foreground main target, the foreground target serif, the speech time of the serif, and the foreground target appearance time. To reproduce the content of the input scenario as an animation, the animation information extraction unit 115 extracts the contents. Animation information alone is not enough. Therefore, in order to complete the animation, it is necessary to add more manual work, and examples of the work items include the following. 1. Set the time to display one frame (hereinafter referred to as playback speed) 2. Search for an image that depicts the substance of the foreground target, and select the image you want to use from the search images 3. Change the utterance time, display time, etc. of the foreground target serifs extracted by the animation information extraction unit 115. 4. Edit the voice quality when the foreground target serifs are uttered. In this embodiment, a method of editing the manner in which the five foreground objects move in the background along with time among the operations of the above items performed by the animation editing unit 120 will be described. As for the other items, the technology for implementing them is known, and therefore the description thereof is omitted here.

Now, in the animation editing unit 120, a method of performing the editing of the movement of the foreground target real image on the background over time, which is the last of the above items, independently of the moving path and the moving speed. Therefore, the editing becomes easy. Hereinafter, a method of editing the manner in which the substantial image of the foreground target moves on the background based on the above method will be described.

Conventionally, the editing of the moving state of the foreground object in animation has been performed as follows. 12 (a) and 12 (b) are conceptual diagrams showing a procedure for editing a moving state of a foreground object by a conventional method.

First, the position of the foreground object at each time is input. In FIG. 12A, points indicated by P1, P2, P3, P4, and P5 are positions where some foreground objects are displayed at times t1, t2, t3, t4, and t5, respectively. Therefore, if the distance between adjacent points is large, the movement between the points will be fast. For example, in the case of FIG. 12 (a), the distance between P3 and P4, P4 and P5 is larger than the distance between other adjacent points, so the movement between the two points is fast. Become.

In order to delay the movement between P3 and P4 in the example of FIG. 12 (a), for example, between P3 and P4, P
To make a new display position of 3.5 foreground, first set P
Prepare a new frame for displaying the foreground object in 3.5 and specify the specific position of P3.5. Then, the time for displaying the foreground object on P3.5 is set to t4, and the time for displaying the foreground object on P4 and P5 is shifted to t4 and t5, respectively. The corrected result is, for example, as shown in FIG.

In the conventional editing method, the moving speed of the foreground object can be adjusted by the above procedure. However, in the method of calculating the moving speed based on the position of the foreground object as described above, in order to adjust the moving speed, while considering the relationship between the time and the display position of the foreground image at the time,
It is necessary to adjust the display position of the foreground object, the number of frames to be inserted or deleted, and its time, which makes it difficult to adjust the moving speed.

In the present embodiment, the above problem is solved by making the means for inputting the movement path of the foreground image and the means for inputting the movement speed at each time independent.

FIG. 13 is a block diagram showing the structure of the animation editing unit 120. Animation editorial department 120
The configuration includes at least a target movement route editing unit 116 that edits the movement route of the foreground target, a target movement distance calculation unit 117, a target movement speed editing unit 118 that edits the movement speed of the foreground target, and a target movement speed calculation unit 119. Have. With this configuration,
Since it is possible to edit the moving path of the foreground object and the moving speed of the foreground object independently, the relationship between the time and the display position of the foreground image at the time is adjusted for adjusting the moving speed. Since it is not necessary to consider, it becomes easy to adjust the moving speed of the foreground object and the display position of the foreground object. Further, in the target moving speed editing unit 118, if the speed at each time is expressed as a ratio, the speed at which the animation is actually displayed is automatically calculated from the display time of the target and the moving path of the target. The editor can edit the moving speed of the target without considering the actual display speed of the target at each time.

Next, the flow of processing in the animation editing unit 120 will be described first, and then the target moving speed editing unit 11 will be described.
The process of 8 will be described in detail.

FIG. 14 is a flow chart showing an example of the flow of processing for editing the manner in which the substantial image of the foreground target moves on the background in time. The flow of processing in the animation editing unit 120 will be described with reference to FIG.

(1401) Input the display time of the foreground object. The input is performed manually by an editor who edits the motion of the foreground target, for example, by the animation editing unit 120.
Is given by the calling module of the call, etc.

(1402) In the target moving route editing section 116, the editor edits the moving route of the foreground target. The editing of the moving path is, for example, drawing the path with a free curve, designating several points as a reference for determining the path, and setting a spline curve connecting the reference points as an actual moving path,
Etc.

(1404) The target moving distance calculation unit 117 calculates the moving distance of the foreground target from the route, that is, the length of the route.

(1405) In the processing of 1401 and 1404, since the display time and the moving distance of the foreground object are fixed respectively, the average speed of the actual movement of the foreground object is calculated.

(1406) The state of movement of the foreground target is checked on the path and speed, and it is determined whether or not to continue editing the foreground target. If not, the editing of the moving state of the foreground target ends.

(1407) If the foreground object is to be edited, then it is determined whether or not the moving path of the foreground object is to be edited.

(1408) When the moving route of the foreground object is edited, the already edited route is corrected.

(1409) Similar to 1404, the moving distance of the foreground object is calculated from the corrected route.

(1410) As a result of the processing of 1407, when the moving path of the foreground object is not edited, it is judged whether or not the moving speed of the foreground object is edited. When the editing is not performed, the moving state of the foreground object is edited. finish.

(1411) When editing the moving speed of the foreground object, the speed already edited is corrected.

(1412) Based on the corrected moving path or the corrected moving speed, the target moving speed editing unit
At 118, the actual moving speed of the foreground object at each time is calculated.

The above steps 1405 to 1412 are repeated until the editor is satisfied, and when satisfied, the editing of the moving state of the foreground object is ended.

Next, the processing in the target moving speed editing unit 118 will be described. FIGS. 15A to 15E are conceptual diagrams showing the target moving speed editing unit 118, and show, as an example, a state in which the speed ratio of each target time is input in a graph format.

In the processing of the animation editing unit 120, when the foreground object is displayed in the processing of 1405 based on the average speed of the foreground object calculated in the processing of 1404, the speed of the foreground object at each time is displayed in a graph format. When expressed with, for example, it becomes as shown in FIG. That is, this indicates that the foreground object moves on the route at a constant speed (FIG. 15B shows that when the foreground object moves on the route at a constant speed, the movement time is divided into 10 parts. It is a conceptual diagram showing the position of the foreground target at each time).

If it is desired to change the calculated moving speed, it is necessary to edit the speed of the foreground object in the processing of 1411.

FIG. 15C shows an example of the edited result of the speed ratio in the form of a graph. The content of the graph shows that the foreground object is stationary at first, increasing the speed by a constant speed, reaching the maximum speed in just half the display time, and then increasing the speed by a constant speed. It drops and then rests again at the end of the display time. In this case, the actual speed of the foreground object at each time can be calculated as follows using the formula of (Equation 1).

[0125]

[Equation 1]

However, V (t) in the equation (1) represents the speed of the foreground object at time t, and T represents the time when the movement ends.

It is assumed that the moving distance of the foreground target calculated in the processing of 1409 is S0. A temporary moving distance when the foreground object moves at the temporary speed edited in the processing of 1411 is calculated by the equation (1), and the moving distance is S1. here,
When calculating the moving distance by the formula (1) based on the speed edited by the processing of 1411, the target moving speed editing unit 118 predetermines the scale of the vertical axis (that is, the speed scale) appropriately. deep.

Next, r is calculated by the equation (2), and the actual moving speed of the foreground object is calculated by multiplying the speed at each time edited in the processing of 1411 by the above r. For example, FIG. 15 (a)
As a result of calculating r using the velocity graph in (c), r = 0.5
Then, the actual moving speed of the foreground object is as shown in FIG. 15 (d) (that is, the speed at each time in FIG. 15 (d) is 0.5 times that in FIG. 15 (c). .Also,
FIG. 15 (e) is a conceptual diagram showing the position of the foreground target at each time when the moving time is divided into 10 when the foreground target moves at the speed shown in the graph of FIG. 15 (d). . ). Thus, in the present invention, the ratio r of S1 and S0 is r
It is possible to calculate the actual speed value for the speed of the foreground object edited by the ratio by introducing a mechanism for normalizing the temporary speed edited by the processing of 1411 using.

As described above, the animation editing unit 120 uses the target moving distance calculating unit 117 by using the speed ratio of the foreground target edited by the target moving speed editing unit 118.
The tentative movement distance calculated by
Since the target moving speed editing unit 118 has introduced a mechanism for calculating the actual value of the speed by using the ratio of the actual moving distance edited in, the animation editor can calculate the actual value of the target at each time. The speed can be edited without considering the moving speed.

In the creation of a normal animation, when the movement of one foreground object appearing in the animation is edited, it is already edited at the same time as the time when the foreground object appears or at a time before or after that time. By referring to the movements of other foreground objects that have already been processed at the same time, it is possible to understand the overall flow of the animation story, and it becomes even easier to edit the desired movement of the foreground object. Therefore, when confirming the movement of the foreground target in the processing of 1405 in FIG. 14, the time when the foreground target appears (1 in FIG.
The time to check to see the movement of the foreground object along with the movement of other foreground objects that have already been edited, including not only the time entered in the processing of 401) but also the time before and after the time if possible. It is considered to be effective to provide the animation editing unit 120 with a function that allows selection of whether or not to simultaneously display the movements of other foreground objects that have already been edited.

In the above, the flow of processing in the animation editing unit 120 for editing the manner in which the foreground target substantial image moves on the background in time with respect to a predetermined time zone has been described.

The animation output control section is based on the information in the animation information storage section 114, and is a sound signal processing section.
While controlling 110, the voice information is sent to the voice synthesizer 109 as information necessary for playing the animation,
The image information is sent to the image display unit 108.

In this embodiment, the animation information storage unit
From the animation information stored in 114, the audio information and the video information necessary for reproducing the animation are extracted, the audio information is sent to the audio synthesis section 109 and the audio signal processing section 110, and the video information is sent to the image display section. I will briefly explain the procedure to play the animation by sending each to 108. As an example, in the scenario (1) above, the control when playing the first movement of "Men's Ken-chan" is
The following is used based on the example of the animation information. -The background image is "at the edge".・ "Mouse no Ken-chan" moves along the edited trajectory while uttering the line "It is not necessary for cats to work for mice" between 0 and 14. The voice quality of the speech is set to [shift the pitch by one octave (hereinafter, this information is referred to as SE1)] with respect to the output sound of the voice synthesis unit 109.

The background image "edge" is first sent to the image display unit 108. Subsequently, each time the frame advances, the foreground image of “Rice's Ken” is sent together with the position and mask information, and the image display unit 108 displays the foreground image on the background image in a superimposed manner. At the time of displaying the foreground image, the mask information is used, and the area where no target is present is made transparent so that the background pixels are displayed.

On the other hand, the voice synthesizing unit 109 has the line "Ken no Ken-chan""Idon't need a cat to work on mice."
However, the voice quality information SE1 of the dialogue is simultaneously sent to the audio signal processing unit 110. The speech is converted into a voice signal by the voice synthesizer 109 and input to the voice signal processor 110. Then, the input signal is converted in the audio signal processing unit 110 on the basis of the information of SE1 and then output as audio by the audio output unit 125.

Next, using the animation creating device,
A procedure for creating data as material used for animation (hereinafter referred to as animation data) and a procedure for creating animation using the animation data will be described. In the explanation of the animation creation procedure, the scenario (1) above is used as the input scenario here as well.

<1. Creation of Animation Data> First, in the image input unit 122, a printed material (for example, a cartoon) on which image data as material for forming each scene of animation is drawn is read, and the read image is It is sent to the image cropping editor.

The image cutout editing unit cuts out only the portion of the image data as the material forming each scene of the animation in the image. A known technique such as a contour extraction technique is used for the cutout.

When the cut out image is used alone, it is sent to the image database management unit 107 as it is, and the image is registered. When using it as an animation brush, create continuous image data that represents the movement of the object using an editing tool, etc. based on the cut out image, group it in order and use it as the animation brush data, and It is sent to the image database management unit 107 and registered. Further, in the case of registration of the background image, the area to be used is sent as it is from the image data read by the image input unit 122 to the image database management unit 107 for registration. Since the editing tool is realized by a known technique such as a model-based image synthesizing technique, the description of the contents of the tool will be omitted. The data registration method in the image database management unit 107 has already been described in the description of the processing in the image database management unit 107, and will be omitted.

<2. Animation Creation> First, the scenario (1) obtained by the scenario input unit 111 is the scenario analysis unit.
It is sent to 113 and the structure of scenario (1) is analyzed. Then, the animation information extraction unit 115 extracts animation information from the structure of the analyzed scenario (1).

FIG. 16 is a conceptual diagram showing a part of the animation information at the time of extraction from scenario (1). In FIG. 16, as the animation information that has already been determined, “ken-chan” as the background target, “ken-chan” as the foreground main target, “ken-chan's mama”, and “ken-chan” as the foreground targets corresponding to the foreground main target. ": Ken no Ken-chan", "Ken-chan dressed as a woman", "Surprising Ken-chan", "Ken-chan's mom": "Ken-chan's mom holding laundry". Further, the dialogue of the foreground target, the utterance time of the dialogue, and the appearance time of the foreground target are as shown in FIG.

Naturally, at this point it is still insufficient to play the animation. Therefore, in the animation editing unit 120, the following two processes are performed manually: correction of information extracted by the animation information extraction unit 115; supplementation of information that cannot be extracted by the animation information extraction unit 115. The details of the processing are as described above.-Set the time for displaying one frame.-Search for an image that depicts the substance of the foreground target and select the image you want to use from the search images.-Animation. Change the utterance time, display time, etc. of the foreground target speech extracted by the information extraction unit 115.Edit the voice quality when the foreground target dialogue is uttered. Edit the way it works with.

FIG. 18A is a conceptual diagram showing a part of the animation information edited in the animation editing unit 120 after being extracted from the scenario (1). Further, the dialog of the foreground target, the vocalization time of the dialog, and the appearance time of the foreground target are as shown in FIG. 18B, for example.
In FIG. 18 (b), the animation editing unit 120 newly adds an image of the foreground target and voice quality of the serif as animation information, and although not shown in FIG. 18 (b), The travel route has also been edited.

When all the animation information necessary for the reproduction of the animation is prepared, finally, the animation information is sent to the animation control unit 121, and the animation control unit 121 outputs the audio signal based on the animation information in the animation information storage unit 114. While controlling the processing unit 110, the audio information is sent to the audio synthesizing unit 109 and the image information is sent to the image display unit 108 as media necessary for reproducing the animation. The voice synthesizing unit 109 converts the input character string into a voice signal, the voice signal processing unit 110 further converts the voice signal, and the voice output unit 125 outputs the converted voice. The image display unit 108 combines and outputs the background image and the foreground image according to the reproduction speed.

The above is the description of the procedure for creating the data as the material used for the animation using the animation creating apparatus of this embodiment and the procedure for creating the animation using the animation data.

[0146]

As described above, according to the present invention, a scenario is analyzed, and from the analysis result, the background object that is the stage background of the animation, the foreground image that appears in the animation, the appearance time of the foreground object, the content of the dialogue, etc. Since it has a function to extract the information required for animation playback,
If the animation creating apparatus of the present invention is used, the animation creator only needs to input a scenario describing the contents of the animation into the apparatus before starting the editing of the animation, and the information for editing the animation can be obtained to some extent. Since it can be acquired automatically, it is possible to reduce the trouble of editing the animation.

Further, in the animation editing unit, the temporary moving distance calculated by the target moving distance calculating unit and the target moving route editing unit are used by using the ratio of the speed of the foreground target edited by the target moving speed editing unit. Since the mechanism for calculating the actual value of the speed is introduced in the target moving speed editing unit by using the ratio of the actual moving distance edited by, the animation editor can change the actual moving speed of the target at each time. It is possible to edit the speed without considering the speed.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an animation creating apparatus according to an embodiment of the present invention.

FIG. 2A is a conceptual diagram showing the contents of image data in a pixel data storage unit of the present embodiment, FIG. 2B is a description example of the same animation brush data storage unit, and FIG. 2C is the same animation brush. Conceptual diagrams (d) and (e) showing the contents of the image are examples of the description of the image search table.

FIG. 3 is a flowchart showing a flow of processing for searching an image in the image database management unit.

FIG. 4 is a flowchart showing a flow of processing for registering an image in the image database management unit.

FIG. 5 is a flowchart showing a flow of processing in the natural language processing unit for extracting nouns corresponding to two of a subject and an action subject.

FIG. 6 is a flowchart showing a flow of processing for extracting a time required for utterance of an input natural language sentence in the natural language processing unit.

FIG. 7 (a) shows a description example of a set of a word heading and a heading representing the pronunciation of the word in the same language dictionary. FIG. 7 (b) shows a heading of the word and the pronunciation of the word in the same language dictionary. Figure showing an example of the description of the time required for

FIG. 8 is a diagram showing an example in which a scenario input to the scenario input section is described in Japanese.

FIG. 9 is a diagram showing an example in which knowledge for analyzing a scenario stored in the scenario analysis dictionary is defined based on BNF notation.

FIG. 10 is a diagram showing an example of a correspondence relationship between terminal symbols and non-terminal symbols and animation information in the knowledge for analyzing a scenario stored in the scenario analysis dictionary.

FIG. 11 (a) is a flowchart showing the outline of the processing in the animation information extracting unit, and FIG. 11 (b) is a flow of processing in the animation information extracting unit for derivatively extracting animation information from the foreground target and the serif of the foreground target. Flowchart showing

[Figure 12] (a) and (b) are
Conceptual diagram showing the procedure for editing the moving state of the foreground object by the conventional method

FIG. 13 is a block diagram showing the configuration of the animation editing unit.

FIG. 14 is a flowchart showing an example of the flow of processing for editing a state in which the foreground target substantial image moves on the background in time.

15 (a) to (e) are conceptual diagrams showing the target moving speed editing unit.

FIG. 16 is a conceptual diagram showing a part of the animation information.

FIG. 17 is a diagram showing a part of the animation information.

FIG. 18 (a) is a conceptual diagram showing a part of the same animation information, and (b) is a diagram showing a part of the same animation information.

[Explanation of symbols]

 101 Language dictionary 102 Natural language processing unit 103 Image data storage unit 104 Anime brush data storage unit 105 Image search table 106 Image database 107 Image database management unit 108 Image display unit 109 Speech synthesis unit 110 Speech signal processing unit 111 Scenario input unit 112 Scenario Analysis dictionary 113 Scenario analysis unit 114 Animation information storage unit 115 Animation information extraction unit 116 Target moving path editing unit 117 Target moving distance calculation unit 118 Target moving speed editing unit 119 Target moving speed calculating unit 120 Animation editing unit 121 Animation control unit 122 Image Input section 123 Feature extraction section 124 Image cutout editing section 125 Audio output section

Claims (23)

[Claims] 1. A language dictionary that stores information necessary for analyzing a natural language, a natural language processing unit that analyzes the natural language using the information stored in the language dictionary, and an image used for animation. An image database that stores the image data, an image database management unit that registers an image used for animation in the image database, and searches the image database; an image display unit that displays the image data; and a natural language input. A voice synthesis unit that converts the natural language into a voice signal, a voice signal processing unit that processes the voice signal, a scenario input unit that inputs a scenario sentence, and a knowledge necessary for analyzing the structure of the scenario sentence are stored. And a scenario analysis unit that analyzes the structure of a scenario sentence using the knowledge stored in the scenario analysis dictionary. An animation information storage unit that stores animation information that is information necessary for playing the animation, an animation information extraction unit that extracts the animation information from the analysis result of the structure of the scenario sentence, and stores the animation information in the animation information storage unit; An animation editing section for editing the animation information stored in the information storage section if necessary and adding new information, and the audio signal based on the animation information stored in the animation information storage section. An animation creating apparatus comprising: an animation control unit that sends voice information to the voice synthesis unit and image information to an image display unit as information necessary for reproducing an animation while controlling the processing unit. 2. An image input unit for inputting an image necessary for creating an animation, and an image cutout editing unit for creating data to be stored in an image database from the image input by the image input unit. The animation creating apparatus according to claim 1. 3. A natural language processing unit analyzes a grammatical structure of the natural language from an input natural language using a language dictionary, refers to a result of analysis of the grammatical structure, and refers to a noun that is a main subject in the input natural language. Is extracted.
The described animation creation device. 4. The language dictionary has means for storing at least a set of a word heading and a time required to pronounce the word, and the natural language processing section uses the information of the set of the language dictionary to store a natural language. The animation creating apparatus according to claim 1, further comprising means for calculating a time required for uttering. 5. The language dictionary has means for storing at least a set of a headline of a word and a headline representing the pronunciation of the word, and the natural language processing unit requires the pronunciation of the word from the headline representing the pronunciation of the word. The animation according to claim 1, further comprising means for calculating a time, and means for calculating a time required for uttering a natural language by using the information of the set of the language dictionary and the means. Creation device. 6. An image database, an image data storage unit for storing image data consisting of attribute information including at least a target attribute and a file name, as pixel data describing information of each pixel and information relating to image content, and an animation. An animation brush data storage unit for storing the animation brush data in which the sets of the image data that make up the above are combined in an ordered manner, and at least three description units of a search keyword, an animation brush data name, and an image data name. The animation creating apparatus according to claim 1, wherein the animation creating apparatus has a configuration including an image search table that stores image search data. 7. The image database comprises an image data storage unit for storing image data including pixel data describing information of each pixel and attribute data including at least a target attribute and a file name as information about image content, and an animation. The animation brush data storage unit that stores the animation brush data in which the set of pointers to the image data is put together in a certain order, and is composed of at least three description parts of the search keyword, the animation brush data name, and the image data name. The animation creating apparatus according to claim 1, wherein the animation creating apparatus has a configuration including an image search table that stores image search data. 8. The animation creating apparatus according to claim 6, wherein the target attribute of the attribute information of the image data has at least two values, that is, a foreground target and a background target. 9. An image database management unit inputs a natural language and refers to a result of analysis of the input natural language by the natural language processing unit to obtain animation brush data or image data related to the input natural language. The animation creating apparatus according to claim 6 or 7, wherein the animation creating database is searched. 10. The image database management unit, when registering the animation brush data in the image search data of the image database, sets the same image data as the image data forming the animation brush data to the image data name of the image search data. 10. The animation creating apparatus according to claim 9, wherein, if it is already described in the description section, the process for deleting the image data described in the description section of the image data name is performed. 11. The image database management unit, when searching for image data from the image database, the image data described in the description section of the image data name of the image search table, and the animation brush data of the image search table. 11. The animation creating apparatus according to claim 10, wherein the image data pointed by a pointer to the image data forming the animation brush data described in the name description section is used as the search result. 12. The animation information storage unit stores at least background information that describes a background object that is a stage background of the animation, at least a foreground object that appears in the animation, a foreground main object that represents the subject of the foreground object, and a foreground object. A foreground information which includes a line, a vocalization time of the line, an appearance time of the foreground target, a moving route of the foreground target, a moving speed of the foreground target, and a voice quality of the foreground target. The animation creating apparatus according to claim 1. 13. The animation creating apparatus according to claim 12, wherein the information on the voice quality of the foreground target of the foreground information is a signal for controlling an audio signal processing unit. 14. The animation information extraction unit extracts a natural language representing a background object which is a stage background of the animation from the analysis result of the scenario analysis unit, and records the natural language as background information of the animation information storage unit. 14. The animation creating device according to claim 12, wherein the animation creating device is a device. 15. An animation information extraction unit extracts at least a set of a natural language representing a foreground target appearing in an animation and a natural language representing a serif of the foreground target from the analysis result of the scenario analysis unit, and extracts the foreground target. 14. The natural language animation information storage unit that represents the foreground object of the foreground information, and the natural language that represents the serif of the foreground object are recorded as the serif of the foreground object of the foreground information, respectively. Animation creation device. 16. An animation information extraction unit extracts, using a natural language processing unit, a noun that is a main subject of the foreground object in the foreground information in which the foreground object is recorded, and extracts the noun as the foreground main object of the foreground object. 16. The animation creating apparatus according to claim 15, wherein the animation is recorded as 17. The animation information extraction unit calculates, for the foreground information in which the foreground target serif is recorded, the time required for utterance of the serif from the natural language representing the serif, using a natural language processing unit, 16. The speech utterance time of the foreground target is calculated with reference to the already calculated serif utterance time of the foreground target, and the calculation result is registered as the utterance time of the serif of the foreground information. Animation creation device. 18. The animation creating apparatus according to claim 17, wherein the animation information extraction unit gives, for the foreground information having no foreground target serif, a time required for utterance of the serif with an appropriate constant value. 19. The foreground information for which the speech time of the speech is registered and the appearance time of the foreground target is not registered, the animation information extraction unit determines the time of speech of the serif of the foreground target by the foreground. 16. The animation creating apparatus according to claim 15, wherein the appearance time of the information is recorded as the appearance time of the foreground object. 20. At least an animation editing unit,
A target moving path editing unit that edits the moving path of the foreground target, a target moving distance calculating unit that calculates the distance of the moving path, a target moving speed editing unit that edits the moving speed of the foreground target, and an actual value of the moving speed. An object moving speed calculation unit for calculating, and records the edited moving path and moving speed as a moving path of the foreground object and a moving speed of the foreground object of the foreground information in which the foreground object is recorded. The animation creating apparatus according to claim 15. 21. The animation creating apparatus according to claim 20, wherein the target moving speed editing unit of the animation editing unit edits the moving speed of the foreground target at a speed ratio. 22. The animation editing unit, other than the foreground information stored in the animation information storage unit, when the information on the voice quality of the foreground target of certain foreground information stored in the animation information storage unit is changed. And, and
14. For the foreground information having the same foreground main target as the foreground main target of the foreground information, the voice quality of the foreground target of the foreground information is changed to the changed voice quality of the foreground target. The described animation creation device. 23. The animation control unit, at the same time as sending the speech content of the foreground target of the foreground information to the voice synthesizing unit,
14. The animation creating apparatus according to claim 13, wherein the information on the voice quality of the foreground target of the foreground information is sent to the audio signal processing unit.