Local Segmentation of Noisy Images

Publisher Website
Google Scholar
Abstract
Images whose properties are spatially variant must often be processed locally. Statistical techniques may be required to do this if the image is noisy. These may be difficult to apply when local regions are so small that means, variances, and similar quantities are unstable. We demonstrate how a practical statistical segmentation algorithm may be constructed which operates locally and gives satisfactory global results. The size of the local area over which computations are made has an important effect on the segmentation quality.

This publication has 0 references indexed in Scilit: