Interactive video data visualization system based on semantic organization

摘要

In this paper, we present a video data visualization system allowing a simple navigation and an effective research in large scale video data corpora. Our tool is based on three steps. The first step consists in the analysis of the video documents: the extraction of high level concepts and the computation of the similarity between the documents of the corpus. The second step is devoted to the preparation of the visualization interface: the projection of the obtained classes in the visualization space and the establishment of the visualization graph. The last step consists to the personalization of the system: the integration of a user model in the visualization system.