Image Processing.
Abstract
This report is organized according to the topics we have worked under this project which include statistical image segmentation, two-dimensional ARMA models, multichannel (multivariate) maximum entropy spectral analysis, and non-Gaussian signal processing, etc. Under each topic, the reports and papers published or presented are also listed. A new AED-512 Imaging/Graphics terminal has been installed in our PDP11/45 minicomputer. The use of the terminal for the image processing project research is also presented. Statistical image segmentation refers to the computer-oriented procedures that partition the image into meaningful parts by using the statistical pattern recognition techniques. A number of techniques have been studied, some of which are supervised while others are unsupervised. A critical evaluation of these techniques has been made. Furthermore we have performed an extensive computer study of the Fisher's linear discriminant method, maximum likelihood estimation and decision-directed method for image segmentation.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 16, 1981
- Accession Number
- ADA095552
Entities
People
- Chia‐Hung Chen
Organizations
- University of Massachusetts Dartmouth