Evaluating Ordinal Pattern Features for 2D Colored Noise Classification

Abstract

This study explores the potential of permutation entropy and statistical complexity for analyzing time series and image data of varying dimensions and noise types to extract features for computational vision. We projected one-dimensional colored noise of different sizes and one-and two-dimensional 1 /f noise with different embedding dimensions to observe changes in permutation entropy and statistical complexity. The results of this study provide insights into the usefulness of the permutation entropy and statistical complexity in the analysis of complex time series data for future parameter extraction