Research On Content Structure Feature Extraction Of Flash Animation

The development of the Internet and multimedia technology promotes the advance of Information Age, which leads to a profound change in the education field. MOOC, micro lessons, mobile learning and other digital learning method are increasingly becoming an important symbol of Education Informatization. Digital learning resources are not only the fundamental support of digital learning, but also its basic elements. The access and application of such resources have a direct impact on the quality and efficiency of digital learning. Flash animation, as the important digital learning resource, has been widely used in the field of education because of its good interaction, simple production process, small-capacity storage space as well as its easy network transmission characteristics.Flash animation can integrate multiple forms of media such as audio, video, text and so on. It also can produce a strong visual appeal by masks, filters, deformation and other treatment. Though the quantity of Flash resource continues to increase,the perfect retrieval system for Flash animation is rare. As a result, the development of retrieval system for Flash based on its content character is quite necessary.The content structure feature of Flash animation, which is the presentation of Flash content character in quantity, plays an important part in its content structure. The extraction results of Flash animation content structure feature can be directly applied to retrieve the Flash animations. It can improve the effectiveness of the Flash animation retrieval system that based on the content structure feature, making retrieval more comprehensive, efficient and accurate, as well as promoting Flash animations to be used in the field of education more widely.The advantages of Flash animation are due to its large number of constituent elements, powerful dynamic effects and perfect interactive features. In this paper, basing on the storage structures and formation principle of Flash animations, we deeply studied the metadata information, the constituent elements, dynamic effects, interactive features of Flash animation from the point of its file format. Besides, four levels of the feature extraction in Flash animation content are put forward at same time. They respectively are the extraction of metadata information, the extraction of the basic elements as well as dynamic effects and interactive features built on top of the basic elements, the extraction of feature scenarios(including logic scenes and visual scene) and semantic feature extraction. We briefly introduce the method of extracting metadata information and specifically demonstrate the description of the constituent elements of Flash animation, dynamic effects, inter-active features and how they are defined as well as extracted. Flash Animation includes some constituent elements such as shapes, images, audios, videos, text, movie clips, shape morphing and buttons, some dynamic effects such as moves, rotates, scales, color changes and shape changes, some interactive features such as interactive buttons, movie clips and so on. On the basis of Flash animation content structure feature, this paper selects 17 types of features, using BP neural network algorithm to realize the automatic classification of Flash animation. The file structure, data types, operating principles and the storage mechanism of Flash animation are used in this paper. We adopt VC6.0 and C++ to achieve the purpose of extracting Flash animation content automatically and to build the content storage as well as indexing database based on the following contents, the basic elements of Flash animation, dynamic effects, interactive features and the extraction method. This paper with considerable innovation and practical applications can contribute to the research and development of content management and retrieval system for Flash animation.