A video annotation methodology for interactive video sequence generation

The FRAMES project within the RDN CRC (Cooperative Research Centre for Research Data Networks) has developed an experimental environment for dynamic virtual video sequence synthesis from databases of video data. A major issue for the development of dynamic interactive video applications of this type is that of how to model and design a video annotation space in order to achieve specific design goals. Consideration of the behavior of the FRAMES algorithm for sequence synthesis, together with the structure of the video annotation space, can provide very clear guidelines for the design of the annotation space, and hence the nature of the video components needed. The resulting design methodology for interactive video is simple to use, and represents a straightforward extension of conventional pre-production activities in standard video production. This paper presents an analysis of the associative chaining algorithm as a search process, together with a systematic video semantic annotation methodology based upon that analysis that can be used to optimise possibilities for user interaction or to achieve specific aesthetic effects. The methodology can be used to create predictable video synthesis behavior, and provides a basis for interactive video production throughout the production lifecycle.