Topological Signal Analysis for Multi-Modal Data Analysis

Extracting information from different modalities of data is one of the greatest challenges in this era of exponential data growth, as well as one of the most promising horizons for the development of new capabilities for military, commercial, and scientific applications. Virtually all areas of defense, industry, and science are becoming increasingly dependent on the ability to analyze multimodal data, and consider the difficulties of this process as one of the most important bot- tlenecks preventing progress. This proposed project addresses one aspect of this challenge by developing methods to detect targets of interest in a variety of situations, using multi-modal sensor data from heterogeneous ISR data sources. We look to not only associate that data, but to do so with appropriate measurements of accuracy in order to use these joint data correctly. Joint distributions on multi-modal data can then provide meaningful probabilities that allow us to deal with conflicts or ambiguities from different sources, and deal more accurately with missing data in one or more sources due to noise, sensor failure or occlusion.