HISTOGRAM COMPARISON USING INTERSECTION METRIC APLLIED TO DIGITAL IMAGES ANALYSIS

Abstract

This paper reports the evaluation results of developed comparison histogram software using the intersection metric. The images used to support our conclusions are from The Amsterdam Library of Object Images, which contains image from objects under different orientations and illuminations. As related in the previous papers, the program indicates that: hue channel color is the least sensitive channel when considering the different object orientations; and the illumination influences all channels when any effective color-constancy algorithm was applied under variable light conditions. Therefore, the developed software seems to be compatible with literature and it is able to be applied in future researches.