Dynamic Topology of Motion Element Based on Sparse Random Graph Grammar

摘要

The dynamic topology structure of motion element in video has led to recent interest in multimedia approaches, which is essential for video coding and understanding. However, the explicit relationship between time and space of it remains unclear. We describe a method in the topology of motion element for the integration of temporal and spatial semantic information by use of semantic space and time. Using sparse random graph, we measure the context of mobile objects and determine the fidelity and structure of the generated state, providing proof of establishment by the statistical semantic information of video motion element. Ultimately, the experimental result is encouraging and indicates that the proposed approach is effective.