Abstract

The need to resize an image is a common occurrence in all facets of industry, from data collection to consumer use. However, most of the research conducted in this area focuses on the various statistical measures, such as RMS deviation. While the statistical measures are suitable for determining the effectiveness of the interpolator for data collection, they do not necessarily coincide with the effectiveness of the interpolator in a consumer application. This study, therefore, focuses on the use of the three most common resizing algorithms, the bi-cubic, bi-linear and nearest neighbor interpolators, in the area of image quality from the consumer application perspective. The study is based on a psychophysical experiment designed to ascertain the preference of a particular algorithm for a particular type of image and determine what characteristics are used to form those preferences. The results are presented so that they may provide direction in the use and development of image interpolators for consumer applications.

Table of Contents