TW201837739A - Animation generating system and method thereof - Google Patents

Abstract

An animation generating system including an input interface, a database, an analysis module and a generating module is provided. The input interface is provided for a user to input multiple words desired to generate an animation. The database is stored in a server and includes multiple sub animations. The analysis module is disposed at the server for analyzing the words and accordingly generating an analysis result. The generating module is disposed at the server and connected with the database and the analysis module. The generating module searches the sub animations from the database according to the analysis result and synthesizes the sub animations corresponding to the analysis result as an animation corresponding to the words. An animation generating method is also provided for a user to input words and then analysis the words and synthesize and generate an animation corresponding the words.

Description

 Animation generation system and method thereof  

The present invention relates to a method and system for generating animation via text, and more particularly to a system and system for automatically generating animations.

Existing dynamic news or teaching systems have animations with captions or subtitles as a means of expression. In other words, the existing dynamic news or teaching system is based on the text content to be conveyed, and is further combined with animation to assist the audience or users in understanding the text content. However, the animation is usually based on completed news content or teaching articles, that is, the producer must first complete the writing of the news content or the teaching article, and then complete the corresponding word by word according to the news content or the teaching article. Animation. It can be seen that the conventional method is costly and time consuming in completing the entire animation.

The animation production system and method of the present invention can automatically and instantaneously generate a desired animation in accordance with input content.

The animation production system of the present invention includes an input interface, a database, an analysis module, and a production module. The input interface is for the user to input a plurality of characters to be animated. The database is stored in a server and includes a plurality of sub-animations. The analysis module is disposed on the server for analyzing text and generating an analysis result accordingly. The generating module is disposed on the server and connected to the database and the analysis module. The generating module searches the sub-animation from the database according to the analysis result, and synthesizes the sub-animation corresponding to the analysis result to correspond to one of the text animations.

The animation generating method of the present invention is for the user to analyze the text on a server and combine to generate an animation of the corresponding text content after inputting the text, which includes the following steps. In a text input step, the user inputs a plurality of characters to be animated via an input interface. In an analysis step, the server analyzes the text and generates an analysis result accordingly. In an animation generating step, the server searches for a stored plurality of sub-animations in a database corresponding to the analysis result, and synthesizes the sub-animation into one of the corresponding texts after obtaining the sub-animation corresponding to the analysis result.

In an embodiment of the invention, wherein the text comprises an article or a subtitle to be used for animation.

In an embodiment of the invention, the text includes at least one subject, at least one verb, at least one accepted word, or at least one scene narrative.

In an embodiment of the invention, the analysis module analyzes and distinguishes at least one main word, at least one verb, at least one accepted word or at least one scene narrative included in the text.

In an embodiment of the invention, the analyzing step includes discriminating at least one subject, at least one verb, at least one accepted word, or at least one scene narrative included in the text.

In an embodiment of the invention, the analysis module includes a parser.

In an embodiment of the invention, wherein the analyzing step comprises analyzing using a grammar analyzer.

In an embodiment of the invention, the database includes a plurality of groups that classify the sub-animation according to different part of speech, and each of the groups is provided with a sub-animation corresponding to the part of speech.

In an embodiment of the present invention, the animation generating method further includes a data establishing step of establishing a plurality of groups in the database for classifying the sub-animation according to different part of speech, and each group has a sub-character. Animation.

In an embodiment of the invention, the group includes at least one subject group, at least one verb group, at least one word group, or at least one scene narration group.

In an embodiment of the invention, the database is connected to a data analysis module. The data analysis module analyzes the plurality of pieces of data, and correspondingly creates the sub-animations and then stores them in the server according to the type of the group.

In an embodiment of the invention, the data establishing step includes connecting the database to a data analysis module.

It can be seen that the animation generating system and the method thereof are adapted to analyze the text and generate an analysis result through the analysis module after the user inputs a plurality of characters to be animated through the input interface, and then pass the analysis result. The generating module searches the stored sub-animation from the database according to the analysis result, and synthesizes the sub-animation of the corresponding analysis result into an animation corresponding to the text. In this way, the animation generating system and the method of the present invention can automatically and instantaneously generate the required animation according to the input content after the user inputs the text according to the requirement, and can be widely used in dynamic news or teaching systems, etc. In the system.

100‧‧‧Animation production system

102‧‧‧Server

110‧‧‧Input interface

120‧‧‧Database

130‧‧‧Analysis module

140‧‧‧ generating module

150‧‧‧Data Analysis Module

S110-S140‧‧‧Steps

1 is a schematic diagram of an animation generation system in accordance with an embodiment of the present invention.

2 is a flow chart of an animation generating method according to an embodiment of the present invention.

1 is a schematic diagram of an animation generation system in accordance with an embodiment of the present invention. Referring to FIG. 1 , in the embodiment, the animation generating system 100 includes an input interface 110 , a database 120 , an analysis module 130 , and a generation module 140 . The input interface 110 allows a user (for example, the aforementioned animator) to input a plurality of characters to be animated. The database 120 is stored in the server 102 and includes a plurality of sub-animations (as detailed below). The analysis module 130 is disposed on the server 102 for analyzing text input by the user and generating an analysis result accordingly. The generating module 140 is disposed on the server 102 and connected to the database 120 and the analysis module 130. The generating module 140 searches for the sub-animation from the database 120 according to the analysis result, and synthesizes the sub-animation corresponding to the analysis result into an animation corresponding to the text.

Specifically, in the embodiment, the text is, for example, an article to be animated or a subtitle to be used for animation, which may be a news content or a teaching term, but the invention is not limited thereto. The user inputs the text through the input interface 110, and the input result can be transmitted to the analysis module 130 disposed on the server 102 by wire or wirelessly. In order to understand what kind of animation is required to be generated, the animation generating system 100 analyzes the text by the analysis module 130 after the text is input. The text includes at least one main word, at least one verb, at least one received word, or at least one scene narrative, and the analyzing refers to analyzing the part of speech of the text, that is, analyzing and analyzing the text by the analysis module 120. At least one subject, at least one verb, at least one accepted word, or at least one scene narrative is included to understand what kind of animation needs to be generated correspondingly.

For example, in the present embodiment, the analysis module 130 includes a parser. The Dependency Parser provided by the Natural Language Processing (NLP) team at Stanford University is taken as an example (the present invention is not limited to the use of this software), at the user input. After the text (for example, an article or a subtitle), the parser as the analysis module 130 extracts the words one by one from the characters, and then selects shift, arc-left, and arc-right according to the relationship between each word and the context. One of the three grammatical actions (ie, determining the part of speech of the word). If the result of the selection of the word is arc-left or arc-right, the connection relationship is further marked. After the processing of all the words is completed, the above selection result will form a parse tree, or a dependency tree. After that, the parser will further find the main words and the accepted words in the sentence.

In English, for example, when the user enters the English "dogs bites me", the analysis result of the parser uses the verb "bites" as the root (root) and the nouns "dogs" and "me" as the roots. The leaves are constructed to construct a parse tree, and then the "dogs" is judged as the grammar "nsubj" (that is, the main word) and the "me" is the grammar "dobj" (that is, the representative words) through the context of the context. Take Chinese as an example. When the user enters Chinese "the victim throws garbage anywhere and accidentally causes murder.", the parser analysis results decompose each word one by one as follows:

Nsubj (lost-3, victim-1)

Advmod (lost-3, anywhere -2)

Root(ROOT-0, lose-3)

Dobj (lost-3, garbage-4)

Punct (throw-3,,-5)

Nsubj (trigger -7, accident-6)

Conj (throw-3, throw -7)

Dobj (trigger -7, killer -8)

Punct (throw-3, .-9)

Thus constructing a dependency tree, in which the grammar of the grammar analyzer "nsubj" represents the subject, so the grammar analyzer judges that "victim" is the subject and the verb is "lost", and the grammar of the parser "dobj" represents the subject, so The parser thus judges that the word "lost" is "junk". Similarly, the parser can use the same method to analyze the relationship between "accident", "trigger" and "killer" in another sentence.

Here are two other two available parsers. First, the parser can use the syntax of "rule-based + POS tagging" to first mark each word in the text with the part of speech, and then find out the noun-verb-noun pattern (NOUN- VERB-NOUN pattern), by which the first noun is the main word and the second noun is the accepted word, and other patterns can be further found, such as the noun-BE verb-verb-introduction BY-noun pattern (NOUN -BE-VERB-BY-NOUN pattern), which means that the first one is the accepted word and the second one is the subject. In Chinese, for example, when the user inputs Chinese "the victim throws garbage anywhere, the accident triggers the killing machine", the grammar analyzer first marks the word in the text as the part of speech, that is, the noun "victim" is the grammar "/NN", the adverb "place" is the grammar "/AD", the verb "lost" is the grammar "/VV", the noun "junk" is the grammar "/NN", the punctuation mark "," is the grammar "/PU", and the noun "accident" is the grammar. "/NN", the verb "trigger" is the grammar "/VV", the noun "killer" is the grammar ""/NN", and the punctuation mark "." is the grammar "/PU". After that, find the pre-set pattern, for example, use the noun-adverb-verb-noun pattern to find the words "victim" - "place" - "lost" - "junk", and use the noun-verb-noun pattern Find the words "accident" - "trigger" - "killing machine", etc., so that you can know the results of analysis such as subject and subject. In addition, the parser can use the syntax of "semantic role labeling" to analyze the semantic meaning of each word in the text, for example, the grammar V represents a verb, the grammar A0 represents a subject, and the grammar A1 represents a direct vocabulary. The grammar A2 represents the accepted words, so that the analysis results of the subject and the received words can be known. Similarly, when the sentence has a scene narrative, the parser can also perform judgment analysis on the part of the word of the word and the text before and after the article. The above is only an illustration of one of the embodiments, and the present invention does not limit the type of grammar analyzer and the applicable language family.

After the analysis module 130 analyzes the text, the generation module 140 of the connection analysis module 130 obtains the analysis result, and searches for the corresponding sub-animation in the connected database 120 accordingly. In other words, before the user inputs the text through the input interface 110 and the analysis module 130 analyzes the text, it is preferable that the database 120 has built in a plurality of sub-animations.

In detail, in the embodiment, the database 120 includes a plurality of groups that classify the sub-animation according to different part of speech, and each of the groups is provided with a sub-animation corresponding to the part of speech. The group is, for example, at least one subject group, at least one verb group, at least one vocabulary group, or at least one scene narration group (but the invention does not limit the genre thereof), and the foregoing sub-animation is completed in advance. It is then stored in the corresponding group according to its part of speech. That is, the subject group stores the sub-animations commonly used for the nouns as the subject words, while the verb group stores the sub-animations corresponding to the common verbs, and so on to other groups. Since the sub-animation corresponding to the part of speech is set in the group of various parts of speech in the database 120, the generating module 140 can search for the corresponding sub-animation in each group of the database 120 according to the analysis result of the analysis module 130.

Further, in the present embodiment, the database 120 is connected to the data analysis module 150. The data analysis module 150 analyzes the plurality of pieces of data and stores them in the server 102 according to the type of the group. The information is, for example, a large amount of news available, but the invention is not limited thereto. By analyzing a large amount of data through the data analysis module 150, it is possible to summarize common characters, scenes, and actions. Then, according to the result of the above analysis and induction, a new plurality of sub-animations can be created correspondingly, or a plurality of sub-animations can be extracted from the existing images, and then the obtained sub-animations are stored in corresponding groups in the database 120. Taking the scene description as an example, a large number of sub-animations about the scene narrative can be built in the scene narrative group in advance, and then after the analysis module analyzes the text, the database is searched according to the action in the analysis result and the article context. And choose the most suitable scene. For example, when the input text is "marathon", the corresponding sub-animation can be a picture of a person running in an outdoor scene, and when the input text is "writing program", the corresponding sub-animation can be searched. It is a screen in which one to several people operate a computer in the office (for illustrative purposes only, the present invention is not limited to the above form). The more data analyzed in advance by the data analysis module 150, the more sub-animations are set in each group in the database 120, and in the subsequent animation process by the animation generation system 100, the generation module 140 searches from the database 120. The motion of the sub-animation corresponding to the entered text is also more precise.

In addition, in this embodiment, after the generating module 140 searches for the corresponding plurality of sub-animations from the database 120 according to the analysis result obtained by analyzing the characters by the analysis module 130, the generating module 140 further synthesizes the corresponding sub-animations. Complete animation. Specifically, after the analysis module 130 analyzes all the words contained in the text to obtain an analysis result, and after the generation module 140 searches for the corresponding sub-animation from the database 120 according to the analysis result, the generation module 140 can follow the The association and timing of the individual words in the text combines the sub-animations into a complete animation. The synthesizing is, for example, superimposing two or more sub-animations together, or sequentially superimposing two or more sub-animations along a time axis, and the present invention does not limit the synthesizing manner. Therefore, the animation performed by the animation generating system 100 is generated based on the text input by the user, and the foregoing analysis, searching, and synthesizing actions are all performed in the server 102, so the animation generated in the embodiment is generated. The system 100 can be widely applied to systems such as dynamic news or teaching systems that use text matching animations to automatically and instantly generate desired animations based on input content.

2 is a flow chart of an animation generating method according to an embodiment of the present invention. Referring to FIG. 1 and FIG. 2, in the embodiment, the animation generating method is used by the user to analyze the text on the server 102 and generate an animation corresponding to the text content after inputting the text, which includes the following steps. First, in the character input step S110, the user inputs a plurality of characters to be animated via the input interface 110. Next, in the analyzing step S120, the server 102 analyzes the text and generates an analysis result accordingly. Finally, in the animation generating step S130, the server 102 searches the database 120 for the stored plurality of sub-animations corresponding to the analysis result, and synthesizes the sub-animation into the corresponding text animation after obtaining the sub-animation corresponding to the analysis result.

Specifically, in this embodiment, the text input by the user in the text input step S110 includes an article or a subtitle to be used for animation, which includes, for example, at least one main word, at least one verb, at least one accepted word, or at least one. The scenario description, and the analyzing step S120 includes, for example, using the analysis module 130 disposed on the server 102 to identify at least one subject, at least one verb, at least one vocabulary, or at least one scene narration included in the text, and includes, for example, using a grammar. The analyzer performs the analysis. However, the invention is not limited thereto. The type of the character input in the character input step S110 and the analysis method used in the analysis step S120 can be adjusted according to the demand.

Furthermore, in the embodiment, the animation generating step S130 is actually included in the server 102 (for example, by using the generating module 140) to search for the stored plurality of sub-animations in the database 120 according to the analysis result. After obtaining the sub-animation corresponding to the analysis result, the sub-animation is synthesized into two steps of the animation of the corresponding text. Therefore, the animation generating method further includes a data establishing step S140, which establishes, in the database 120, a plurality of groups for classifying the sub-animation according to different part of speech (for example, including at least one subject group, at least one verb group, At least one word group or at least one scene description group, and each of the groups is provided with a sub-animation corresponding to its part of speech.

In detail, in the embodiment, the data establishing step S140 includes connecting the database 120 to the data analysis module 150. The data analysis module 150 analyzes the plurality of pieces of data, and stores them in the server 102 according to the type of the group after the sub-animation is created. The means for establishing the data can refer to the foregoing content, for example, analyzing a large amount of existing news to summarize common characters, scenes, and actions, and accordingly, to create a new variety of sub-animations, or to view images from existing ones. A plurality of sub-animations are taken out and then stored in the corresponding group in the database 120. Therefore, the material creation step S140 should be before the animation generation step S130. Preferably, the data establishing step S140 can be performed before the text input step S110, that is, in the first step of the animation generating method, so that the subsequent searching operation in the animation generating step S130 is faster and more accurate. However, the present invention does not limit the above-described order, and it can be adjusted according to requirements, as long as there is a corresponding condition before each step is executed (for example, when the search operation is performed in the animation generating step S130, the database 120 has been stored, for example, through the material establishing step S140. Multi-seed animation).

In summary, the animation generating system and the method thereof are adapted to pass through an analysis mode after a user inputs a plurality of characters (for example, an article or a subtitle to be used for animation) through an input interface. The group analyzes the text and generates an analysis result, and then searches the stored sub-animation from the database according to the analysis result through the generation module, and synthesizes the sub-animation corresponding to the analysis result into an animation corresponding to the text. The sub-pictures stored in the database may be obtained by analyzing the plurality of pieces of data (for example, a large amount of existing news) in advance, and then sub-animation is performed according to the analysis result or the sub-animation is extracted from the existing animation. In this way, the animation generating system and the method of the present invention can automatically and instantaneously generate the required animation according to the input content after the user inputs the text according to the requirement, and can be widely used in dynamic news or teaching systems, etc. In the system.

Claims (16)

 An animation generating system includes: an input interface for a user to input a plurality of characters to be animated; a database stored in a server and including a plurality of sub-animations; and an analysis module disposed on the servo And generating a module, the generating module is disposed on the server, and is connected to the database and the analysis module, and the generating module is configured according to the analysis result Searching for the sub-animations from the database, and synthesizing the sub-animations corresponding to the analysis results corresponding to one of the texts.    The animation production system of claim 1, wherein the text comprises an article or a subtitle to be used for the animation.    The animation production system of claim 1, wherein the characters include at least one subject, at least one verb, at least one vocabulary or at least one scene narration.    The animation generating system of claim 1, wherein the analyzing module analyzes and distinguishes at least one subject, at least one verb, at least one accepted word or at least one scene narrative included in the characters.    The animation production system of claim 1, wherein the analysis module comprises a parser.    The animation generating system of claim 1, wherein the database comprises a plurality of groups for classifying the sub-animations according to different part of speech, and each of the groups is provided with corresponding words. Sub animation.    The animation generating system of claim 6, wherein the groups comprise at least one subject group, at least one verb group, at least one word group, or at least one scene narration group.    The animation generating system of claim 7, wherein the database is connected to a data analysis module, and the data analysis module analyzes the plurality of data, and correspondingly creates the sub-animations according to the groups. Types are classified and stored in the server.    An animation generating method is provided for a user to analyze text and combine to generate an animation of a corresponding text content after inputting a text, and the method includes the following steps: a text input step: the user inputs an plural number of animations to be generated via an input interface. Text analysis; an analysis step: the server analyzes the text and generates an analysis result; and an animation generation step: the server searches for the stored result in a database corresponding to the analysis result Animating, and after obtaining the sub-animations corresponding to the analysis results, synthesizing the sub-animations into one of the texts.    The animation generating method of claim 9, wherein the text includes an article or a subtitle to be used for the animation.    The animation generating method according to claim 9, wherein the characters include at least one main word, at least one verb, at least one accepted word, or at least one scene description.    The method for generating animation according to claim 9, wherein the analyzing step comprises discriminating at least one subject, at least one verb, at least one accepted word or at least one scene narrative included in the characters.    The animation generating method of claim 9, wherein the analyzing step comprises analyzing using a parser.    The method for generating an animation according to claim 9 further includes: a data establishing step of establishing, in the database, a plurality of groups for classifying the sub-animations according to different part of speech, and within the groups Each of these sub-animations corresponding to their part of speech is provided.    The method for generating animation according to claim 14, wherein the groups comprise at least one group of subjects, at least one group of verbs, at least one group of words, or at least one group of scene descriptions.    The method for generating an animation according to claim 15, wherein the data establishing step comprises connecting the database to a data analysis module, wherein the data analysis module analyzes the plurality of data and correspondingly creates the sub-animations. The categories are stored in the server according to the types of the groups.