The Alternative FO-FIRLP-AF System for Eradication of Fix Value Impulsive Noise in Computer Representations

This philosophical article proposes the alternative First-Order Finite Impulse Response Linear Predictor with Average Filter (FO-FIRLP-AF) system, which is researched from the original First-Order Finite Impulse Response Linear Predictor with Median Filter (FO-FIRLP-SMF) system. The earliest track of the noise eradication system is to distinguish whether the inspected representation dot is the noisy representation dot or the original representation dot by the maximum value or the minimum value. In the second track, only noisy representation dots are replaced by the statistical average of the 1st order FIR linear predictor. In the computer verification, multitudinous noisy computer representations that engross of LENA and PEPPER at all noisy percentage are prosecuted by the alternative FO-FIRLP-AF system in contrariety of the productivity of prior state-of-art systems: SMF, MF, AMF, WMF and original FO-FIRLP-SMF system. Due to the verification outcomes, the productivity of the proposed alternative FO-FIRLP-AF system is better than previous prior state-of-art systems in objective assessment and subjective assessment at all noisy percentage, especially low and medium noisy percentage.