Geometric and Topological Methods for Multi-Modal Data Analysis and Fusion

Abstract

We propose research on the use of geometric and topological methods to capture shape in dynamic data, with envisioned applications to multi-modal data fusion, human and vehicle tracking, and activity recognition. We will use data shape in supervised learning, novelty detection, and multi-modal data analysis, emphasizing what’s geometric in the data with minimal use of predetermined models. Modalities that we propose to analyze include, but are not limited to, Signals Intelligence (SIGINT), Ground Moving Target Indicator (GMTI) radar, full-motion video (FMV), and motion-capture (MOCAP). In this multi-modal data analysis, we seek to both improve characterization and discrimination within a single data type as well as manage and utilize information from different data types.

Document Details

Document Type
DoD Grant Award
Publication Date
Jul 11, 2018
Source ID
FA95501810266

Entities

People

  • John Harer

Organizations

  • Air Force Office of Scientific Research
  • Duke University
  • United States Air Force

Tags

Readers

  • Computer Vision.
  • Neural Network Machine Learning.
  • Sensor Fusion and Tracking Systems.

Technology Areas

  • AI & ML