Resolution-Dependent Information Measures for Image Analysis

A picture processing scheme which quantifies pictorial information through the use of a resolution-dependent feature extraction process is introduced. The proposed quantification method (based on an information theoretical approach) can be used for edge detection, texture analysis, and classification, as well as feature extraction. These applications are possible because the information measures obtained are capable of uncovering some basic characteristics of images. Vector quantities and synthesized plots of various information measures derived from images are included here to demonstrate the usefulness of the method in image analysis.