Telecommunications and Radio Engineering

AN APPROACH TO PREDICTION OF SIGNAL-DEPENDENT NOISE REMOVAL EFFICIENCY BY DCT-BASED FILTER

ABSTRACT

An approach to prediction of denoising efficiency for DCT-based filter applied to images corrupted by signal-dependent noise is presented. This approach allows estimating quantitative criteria of filtering efficiency from one statistical parameter that can be quickly calculated for a given noisy image under condition that parameters of signal-dependent noise are a priori known or pre-estimated with appropriate accuracy. We demonstrate in this paper that the prediction approach is applicable to different types of signal-dependent noise. Besides, we show that the statistical parameter used for prediction can be calculated in different ways and this influences prediction accuracy.