Animation-oriented Field Of Ontology Construction And Reasoning

As an important research branch of the narrative intelligent field, automaticgeneration of computer animation have a combination of existing research results ofmany disciplines. It can be widely used for work, daily life and entertainment. Itsrepresentative research is reflected in the natural language of the story to thetranslation of the computer animation.Regarding the automatic generation of computer animation as our researchbackground, we researched an in-depth study of ontology construction and massiveontology of data reasoning, and discussed the animation-based ontology model andRDFS extended reasoning based on the MapReduce. The main works are summarizedas following:Primarily, we analyzed and summarized the research background and currentsituation of automatic generation of computer animation, ontology construction andreasoning. In addition we introduced the theoretical knowledge of ontology, includingthe concept of ontology, ontology modeling, ontology languages, ontology reasoning,and distributed computing theoretical knowledge.Next, we put forward an animation-oriented ontology model through the demandanalysis for animation ontology and the combination of the research of existing basisontology. This model defines the concept of the event, space, time, objects andinformation entities, and build the classify event model refers to the key action in theanimation. And it is established all kinds of relationships between events, events andother objects, other objects based on the event model, emerged the event-basedanimation ontology relational schema. It also used the description logic to express itsaxiomatic. According to the manually instantiate test results, it shows that theontology model have a good description for the animation script, and it ensured thescalability and reusability of the ontology.Finally, we propose an extending reasoning method based on the MapReduceRDFS. This reasoning method using the idea of MapReduce programming model, andHadoop framework for the experimental platform, expanded and implemented a RDFSrules distributed inference algorithm. Firstly this reasoning program will processesthe massive ontology of data stored in the HDFS by compression coding. And then wefollowed by running inference algorithms of various rules according to a certain order of reasoning. And finally decompress the output. Through the building of a smallHadoop cluster program test, the results show that the inference algorithm can notonly significantly improve the efficiency of reasoning, but also have a goodscalability.