Masters Theses

Date of Award

12-1994

Degree Type

Thesis

Degree Name

Master of Science

Major

Computer Science

Major Professor

Bruce Whitehead

Committee Members

Dinesh Mehta, Jack Hansen

Abstract

Image classification can be performed on the basis of textural differences. Many textural techniques have been developed to operate on monochrome images. Color images may contain texture patterns that are dependent on color variations for textural discrimination. When a monochrome textural technique is applied to the gray-tone levels of these color images, the texture patterns may not be distinguishable. Therefore, a color texture energy measure technique was developed based on a monochrome technique developed by Kenneth. I. Laws in 1980. Color texture energy is measured by convolving each RGB plane of a color image separately with a set of zero-sum two-dimensional masks which respond to specific textural features. The results are combined to calculate the color texture energy measures for each mask. The color texture energy measures can then be used for image classification. Accuracies of 88.5 percent with color texture energy measures versus 72.0 percent with monochrome texture energy measures were achieved in this study when each technique was used to classify complex natural color textures.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS