Multi-Level Fusion: issues in bridging the gap between high and low level fusion
Wolfgang Koch, Kellyn Rein, Fraunhofer Fkie, Sten F. Andler, Jesús Garcı́a · 2012
In this digital age, information is continuously flowing in from a multitude of heterogeneous sources, both device-based (sensors, video, etc.) and human-generated (text, voice, etc.). There is a growing need to effectively identify relevant information from the mass available, and exploit it through automatic fusion for timely, comprehensive and accurate situation awareness. The processing of sensor-produced data is now a proven, mature technology and great advances have been made in analysis of text-based information based upon computational linguistics. However, up until now researchers in these two communities have operated within their respective communities with little cross-pollination of ideas, experience and methodologies. As a result, timely, comprehensive and precise situation awareness is impossible because of this communication gap. The lack of symbiosis between these communities is caused in part by the lack of an integrative, holistic approach to combining sensor-based data (low level fusion) and text-based information (high-level fusion). The proposed Panel discussion will bring together researchers from both fusion communities to identify the commonalities and differences in their research methodologies. Keywords-Heterogeneous multi-level information/data fusion, situation awareness, threat assessment, information reliability, context exploitation I. PANEL MOTIVATION