Event-based Plot Planning Of Mobile Animation

The technique of the Full-Life Cycle Automatic Animation Generation is proposed by the Chinese Academy of Sciences Institute of Lu Ruqian academician in the 1990 s.It is based on computer graphics,the integration of artificial intelligence technology and film art,to convert the story of the text to the video video automatically.With the continuous development of mobile communication technology,in order to change the presentation mode of SMS text,the automatic animation generation system of mobile phone short message transformes the opening Chinese natural language text into animation to enhance the attractiveness of SMS,which applys the technology of Full-Life Cycle Automatic Animation Generation.Plot planning is an important part of the automatic animation generation system,which plays an important role in guiding the final animated content.Its main task is to develop abstract events into a series of specific actions that make it possible to describe a story.The evaluation criteria of the plot planning is based on whether the content of the animated plot is able to express the text message,while the richness of the plot and the diversity is one of the important criteria for the evaluation of the plot.In this paper,we explore the plot planning method in the narrative field,study the raleted content of the event concept,and then put forward the event-based plot planning method in the mobile SMS automatic generation system.This method is based on the event ontology library.It will get the key information from the result of natural language processing,and obtain the plot information of the message conforming to the text through semantic query.It will realize the transition from single action to multiple actions,enrich the animation plot content,increase the diversity of the plot,and improve the accuracy of the performance of the animation plot.The main contents of this paper include two aspects:First,the ontology library is created to support the plot planning.Event-based plot planning is based on the ontology library to support the entire planning process,mainly including the event ontology library and action ontology library.In the process of establishing the ontology library,firstly,it is clear that the connection between the plot,the event,and the action,so that the content of the plot can be expressed by using the underlying action data.Secondly,we make clear of the definition of the event and the relationship between events and events in the planning process.Third,the main action in the animation can be used as the reference model to establish the event class,and the event ontology relationship model is formed by establishing the relationship of events and events,events and actions,events and themes,events and templates.And the descriptive logic method is used to axiomatically express the various relations of the event ontology.The event ontology model can show a variety of plot,and provide the basic guarantee for the plot planning.Second,we realize the qualitative planning and quantitative calculation based on the ontology library.The process of qualitative planning is based on the result of information extraction which includes the key information in the text of the message,including the topic,character template,action template and mood template.We query the event ontology library according to the extracted key information,to find the corresponding event and action based on the semantic relationship model.As the topic-event is a many to many relationship,the template-event is a many to many relationship,event-action is a many to many relationship,the resulting plot will be varied through different event combinations,different actions match,then based on random selection strategy,we will select a plot as a qualitative output.In the quantitative calculation process,the animation of the characters is dynamically set by calling the three-dimensional animation software Maya API and scripting languages Mel.By creating the character role object,calculating the start time and the ending time of the action execution,the selected actions are combined in series,so that the action executed by characters can achieve a smooth transition to improve the mobility of the action.We have a total of 968 experimental data in our statistical system over the past few years to verify the feasibility,diversity and ornamental nature of the event-based plot planning method.The number of SMS messages generated by the event-based plot planning method is 312,which verifies that the method can be effectively applied to mobile phone SMS automatic animation generation system.One of the reasons for small number of animations generated is that there is no topic or action template in the information extraction,resulting in no choice of the associated event combination.On the other hand,there is no event corresponding to the topic or action template which the information is extracted.For the analysis of 312 text messages,the number of episodes generated in the animation is at least one,up to 16,indicating that the event-based plot planning method can produce a variety of plot.The experimental data shows that the effect of the animation plot has been better,the situation of simple plot has been improved,and the animation content has been enriched since the method was applied to the system.In order to make the content of the animation be understood by the public,the questionnaire was adopted to test the readability of the event-based plot planning method.Participants evaluate the animation content according to their own subjective understanding.The final data shows that the public satisfaction with the animated plot reached 69.44%,indicating that the animation can be recognized by the public.Based on the automatic animation generation system of mobile phone,this paper puts forward the plot planning method based on event aimed at improving the diversity and richness of animation plot.In order to make the plot no longer limited to a single action and show a more reasonable animation content through a number of actions in series,we start from the "event" and carry out the three aspects of "plot-event-action" planning to build a variety of circumstances through the combination of different events.Further work includes extending the event ontology library which could be through machine learning and statistical methods to obtain the relationship between events to improve the coverage of the event ontology library,while considering the plot planning when the people is not in the scene.