Color Image Restoration Using Iterative Mead’s Filter

Abstract

Image restoration is reconstructing the true image starting from degraded (blurred and noisy) image version. This problem could be handled as blind or non-blind mode depending on whether functional knowledge or point spread function (PSF) knowledge is available. These knowledge are related to the type and parameters of the additive noise such as; distribution, mean, and variance. Accordingly, the present work aims to restore the original image using adaptive Mead’s algorithm applied on degraded version. The proposed method is individually applied on the three color components; red, green, and blue (RGB) of the image. The results are quantitatively and qualitatively compared to the original one using two quality measures, they are: mean squared error (MSE) and cross correlation coefficient (CCC). Results showed valued performance of the proposed method when restoring the degraded images. Quality measures proved that the blue component was reconstructed better than the red, and the red was reconstructed better than the green component. Frequent tests showed the matching score between the reconstructed image and the original one was about 97%, which ensure the validation of the proposed method and correct path of computations.