Efficient Dimensionality Reduction for Massive Image Datasets
Abstract
The goal of the research is to mine low-dimensional geometric structures that are often present in image, video and hyperspectral collections to provide a class of dimensionality reduction techniques that tailor pairwise distances between data points such that specific criteria are preserved.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Oct 01, 2018
- Source ID
- HM04761712000
Entities
People
- Aswin C. Sankaranarayanan
Organizations
- Massachusetts Institute of Technology
- National Geospatial-Intelligence Agency