Objective perceptual evaluation of halftoning using image quality metrics

Abstract

Many color and gray-tone image quality metrics have been developed during the last decades. Unfortunately, these metrics could not be directly applied to halftone images. They failed to reliably estimate the distortion resulting from the halftoning process. Here, we propose to adapt some well-known image quality metrics to evaluate halftone images. One of the most important and primary characteristics of the Human Visual System is its optical lowpass filtering effect which transforms the halftoned images into perceptually continuous images. Therefore, we focus our study on the design of the HVS lowpass filter. It is demonstrated through theoretical developments and experimental studies that the Gaussian filter is a good candidate for mimicking the perceptual lowpass filter. It is then shown that by incorporating this filter in the design of some image quality metrics, a comprehensive and relevant comparison of halftoning techniques could be achieved.