ICA Feature Extraction and SVM Classification of FLIR Imagery
Abstract
Detection and clutter-rejection algorithms are being developed to process forward-looking infrared (FLIR) imagery. Feature selection is an important issue, with principal components analysis (PCA) and independent components analysis (ICA) constituting two algorithms of interest. With regard to processing the features, we are examining Bayesian learning-machine algorithms, such as the relevance-vector machine. All algorithms are applied to measured data.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 15, 2005
- Accession Number
- ADA441506
Entities
People
- Lawrence Carin
Organizations
- Duke University