Adaptive Nonlinear Filter using Main Texture Direction for Mixed Noise in Color Image Processing

Abstract

Adaptive nonlinear filter (ANF) using main texture direction is proposed for the removal of mixed noise from corrupted color images processing. The purpose of this work is to focus on designing noise detector and noise filter. For the detector, Chebyshev’s theorem and the fuzzy mean process are used to estimate adaptive parameters of detector; for the filter, the authors use the local texture direction probability density distributions which are gained by performing the Radon transform of the image. Extensive experimental results show that the proposed ANF outperforms other filters in terms of important evaluation metrics; in particular, the computational complexity of the ANF is much lower than the test filters.