Study of Features of Binary Images Using Algebra Techniques
Final technical rept. 1 Dec 89-30 Nov 90,
NEBRASKA UNIV LINCOLN DEPT OF COMPUTER SCIENCE AND ENGINEERING
Pagination or Media Count:
The aim of this project was to investigate features of binary images by considering a special case of the Image Algebra methodology obtained by representing digitized images both monochrome and colored by certain polynomials in two variables with coefficients from the binary field. Since polynomials can be easily manipulated and our proposed operators can be described conveniently in terms of algebraic operations on these polynomials, this approach provides a significant foundation for practical applications which would be of significant interest to AFOSR. Our specific objectives have been as follows. We have developed algebraic operators in the context of the polynomial approach to determine the contour, magnification and shrinking, and a sequence of approximations from finer to coarser of a binary image. Further, we have extended our techniques to process gray images and do operations such as template decomposition, shape decomposition, connected component labelling. Also, we have developed an algebraic system to process colored images. Also, we have developed some fast sequential and parallel thinning algorithms. We have also extended the polynomial approach to three dimensional by developing equivalents of the standard morphological operations and applying these to do a number of operations for the understanding of three dimensional objects.
- Numerical Mathematics