Spatiotemporal lifelines in support of video queries

In this paper we address the use of lifelines to support spatiotemporal queries. Spatiotemporal lifelines express the behavior of moving objects in a video segment. Accordingly, their analysis offers a powerful abstraction mechanism for the creation of brief summaries of video datasets. Here we present similarity metrics to compare lifelines. Two types of lifeline comparison can be supported First, we can compare lifelines considering only the geometry of the trajectories. Alternatively, we can compare them considering the values of various attributes that express specific object properties like acceleration and orientation. The paper includes some early experimental results to demonstrate the use of the presented metrics