Impulse noise removal using fuzzy logic and alpha-trimmed mean

In this paper, fuzzy logic and alpha-trimmed mean based filter has been proposed for the smooth out uniform impulse noise from corrupted gray-scale images. The proposed method employs the concept of alpha-trimmed mean in combination with fuzzy logic to avoid the outlier effect. Alpha-trimmed mean and median values play an important role to formulate the fuzzy membership function. This function is then finally used to compute the estimated value of the noisy pixel. Experimental results demonstrates that the technique presented in this paper gives better performance than the other impulse noise removal techniques based on well known performance measure namely peak-signal-to-noise-ratio (PSNR) and structural similarity index measure (SSIM).