You can view the full text of this article for free using the link below.

Title: Nonlinear tensor diffusion filter for the denoising of CT/MR images

Authors: S.N. Kumar; A. Lenin Fred; H. Ajay Kumar; P. Sebastin Varghese

Addresses: Department of Electronics and Communication Engineering, Sathyabama Institute of Science and Technology, Chennai – 600119, Tamilnadu, India ' School of Computer Science Engineering, Mar Ephraem College of Engineering and Technology, Elavuvilai – 629171, Marthandam, Tamilnadu, India ' School of Electronics and Communication Engineering, Mar Ephraem College of Engineering and Technology, Elavuvilai – 629171, Marthandam, Tamilnadu, India ' Metro Scans and Laboratory, Thiruvananthapuram, Kerala – 695011, India

Abstract: The partial differential equation based algorithms play a prominent role in image processing and computer vision applications. The anisotropic diffusion technique was widely used for image enhancement and denoising. The Perona-Malika algorithm based on anisotropic diffusion fails to preserve sharp edges and fine details in the denoised image. In this paper, the variants of Perona-Malika (PM) model, nonlinear scalar diffusion (NLSD) filter and nonlinear tensor (NLTD) filter are analysed. The algorithms are analysed on sheep phantom image corrupted with Gaussian and Rician noise and results were validated by performance metrics like PSNR, MAE, EPI and MSSIM. The NLTD filter produces superior results when compared with NLSD and PM filter. The NLTD filter was also found to yield efficient restoration results for real-time CT/MR images and was validated by entropy measure.

Keywords: denoising; Perona-Malika; nonlinear scalar diffusion; NLSD; nonlinear tensor diffusion; NLTD; Rician noise; gauss gradient operator.

DOI: 10.1504/IJAIP.2023.128079

International Journal of Advanced Intelligence Paradigms, 2023 Vol.24 No.1/2, pp.156 - 172

Accepted: 22 Oct 2018
Published online: 05 Jan 2023 *

Full-text access for editors Full-text access for subscribers Free access Comment on this article