Enhancement of sub-sampled chrominance image data

Abstract

In image communication systems, sub-sampling can be used to decrease the transmission rate of images. Since the Human Visual System is less sensitive to chrominance information, one common technique is to sub-sample chrominance information, leaving a high-resolution luminance plane. This process however creates artifacts and results in a degraded image which often contains false colors. Much better visual results can be obtained by enhancing the image after transmission. This paper develops an algorithm for enhancing color image data that has sub-sampled chrominance planes. The algorithm is based on stochastic regularization using the high-resolution luminance channel to guide the enhancement of the sub-sampled channels. Visual results are presented.