Removal of ‘Salt & Pepper’ noise from color images using adaptive fuzzy technique based on histogram estimation

This paper presents an algorithm for removal of ‘salt and pepper’ noise from color images. Adaptive fuzzy filter based on histogram estimation (AFHE) has been proposed in which the size of the processing window is adapted based on local noise densities using fuzzy based criterion. In the algorithm, AFHE has been used iteratively and an additional Gaussian mean based iterative procedure has been incorporated for processing at mid and high density impulse noise corrupted image respectively. The experiments have been carried out on a large database for different classes of images and the performance is measured in terms of PSNR, SSIM and FSIMC. The proposed algorithm outperforms some of the existing state-of-the-art filters in terms of PSNR by at least 6 dB and in terms of the structural similarity by at least 0.3 over 50% noise density. The performance improvement is obtained keeping a trade-off between the restored image quality and the computational complexity. The visual observation, shown in this work, also suggests that the proposed filter provides satisfactory performance even when the image is corrupted at a high impulse noise of 90%.