Abstract
The denoising procedure attenuates the image noise while preserving its edges and fine details. In computed tomography (CT), images are degraded by additive white Gaussian noise because of different acquisition and system errors. Due to noise existence, specialists may encounter certain difficulties to analyse or extract the useful information from noisy images. This article presents a novel implementation of the phase-preserving algorithm to denoise CT images. The phase preserving is a powerful noise reduction algorithm, but it tends to remove specific details from the processed images supposing them as noise. Therefore, a Wiener filter that uses 2D Gaussian point spread function is used along with a modified version of the latter algorithm to reduce the noise and conserve the minor medical details. The performance of the proposed approach is assessed on naturally and synthetically degraded CT images using the universal image quality indexand peak signal-to-noise ratio accuracy metrics. Results show major improvement not only in noise attenuation but also in preserving the small details.