Modified Pixels Based Fast Median Filter in Impulse Noise Environments
10.11591/ijeecs.v9.i3.pp731-741
Man Hee Lee
,
Soo Young Shin
This paper proposes a modified pixel-based fast median filter (MP-FMF) for impulse noise environments. The key idea behind MP-FMF is the reduction in the processing time by using modified pixels. It consists of three steps, namely, error detection, threshold decision, and noise mitigation. The presence of noise is detected by using modified pixels that include a cyclic redundancy check (CRC) function. Subsequently, the threshold values are decided by estimating the noise density. For noise mitigation, corrupted pixels are corrected with the neighboring pixels based on the principle of the median filter. The MP-FMF has a fast processing time and provides image quality correction and introduces features when the noise density is high. In addition, we introduce a new evaluation metric and investigate the performance of the proposed algorithm in terms of the quality, features, and computation time.
