Restoration of Noise Degraded Image by Spectral Subtraction

Abstract

We applied two restoration systems to noise degraded images; one is the spectral subtraction image restoration (SSIR) system and the other is our modified version. In some restoration system such as Wiener filtering, it is necessary to estimate true image power spectrum. However, these systems are unnecessary to estimate it. In this paper, it is shown by some examples how noise degraded images are improved by the change of parameters. At the suitable parameter selected, the results show that these methods give similar improvement comparing with Wiener filtering technique in root-mean-square error.