Automated Analysis of Imagery.
Abstract
The objective of this research was to study possible techniques for extracting features and textural regions from imagery using Generalized Balanced Ternary. In addition, various aggregate measures were defined and their properties investigated as to their applicability in representing features. A mathematical definition of texture was derived based on the variance of the picture function over areas. This definition was tested on sample textures. Among the aggregate measures that were derived and studied was a function, f squared, which distinguishes shapes. f squared is invariant under size, rotation, translation, and scaling in any axis. This measure and others were used in software which converted images in pixel form to higher level representations. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1982
- Accession Number
- ADA125710
Entities
People
- Dean Lucas
- Laurie Gibson