A New Adaptive Alpha-Trimmed Mean Filtering Approach

This paper presents a novel algorithm to restore the images corrupted by mixed noises like the combination of Gaussian noise and impulsive noise. It is based on the alpha-trimmed mean filter and uses symmetric specialty of noise model. Comparison of experiments´ results with the adaptive alpha-trimmed mean filter using asymptotic variance minimization shows that the proposed method has a comparable performance but with approximately twice filtering speed, easiness to implement and better robustness for the mixed noise model deviated from symmetric noise distribution