Survey of Texture Segmentation, Classification, and Synthesis Methods
Abstract
This report reviews the literature in the areas of image texture segmentation, classification, and synthesis methods. The approaches to these areas are grouped into areas of fractal, spline, neural networks, modeling, and stochastic methods. An immense amount of literature was reviewed, and techniques with the most merit are presented. From the review it appears that no single approach provides a robust texture analysis methodology without requiring overwhelming complexity. It seems that the best approach to the problem may be to use a variety of these methods in the simplest and most computationally economic forms.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1992
- Accession Number
- ADA252623
Entities
People
- Brian S. Bourgeois
- Charles L. Walker
Organizations
- United States Naval Research Laboratory