Denoising Accuracy of Adaptive ICI-Based Estimators With Regards to Sampling Rate

Abstract

This paper presents study on denoising accuracy of adaptive temporal filtering methods based on the intersection of confidence intervals (ICI) rule and relative intersection of confidence intervals (RICI) rule with regards to signal sampling rate. The original ICI-based and the improved RICI-based method were tested on four signal classes for a range of signal to noise ratios (SNRs). Denoising accuracy, with respect to signal sampling rate, was measured in terms of the reductions in root mean squared error (RMSE) and mean absolute error (MAE). Extensive simulations showed that the data-driven RICI method outperformed the original ICI method reducing the RSME by up 79.6% and the MAE by up to 86.1%. It is important to note that both methods, especially the RICI method, exhibit significant estimation accuracy improvement in case of signals with higher sampling rates.