Multiscale Path Metrics for the Analysis of Discrete Geometric Structures
Abstract
The objective of this research is to model the topological and combinatorial structure of realistic data such as video that captures moving objects (pedestrians, vehicles, marine vessels) in the context of recognition, tracking, and classification tasks. Each data item (for instance, an image of a handwritten digit, or the bounding box around the image of an observation target in one video frame) is viewed as a point in an abstract space, and the main thrust of this research is to model the topology, geometry, and combinatorics of the sets that result from collecting many points like these from data recordings.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 30, 2017
- Accession Number
- AD1051617
Entities
People
- Carlo Tomasi
Organizations
- Duke University