A Comparison of Spatial Pattern Spectra

Download PDF
Open Webpage

Sander Land, Michael H. F. Wilkinson

Published: 2009, Last Modified: 05 Jan 2026ISMM 2009EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Pattern spectra have frequently been used in image analysis. A drawback is that they are not sensitive to changes in spatial distribution of features. Various methods have been proposed to address this problem. In this paper we compare several of these on both texture classification and image retrieval. Results show that Size Density Spectra are most versatile, and least sensitive to parameter settings.