Manuscript received August 6, 2024; revised September 4, 2024; accepted September 26, 2024; published March 14, 2025.

Abstract—The critical objective of image denoising is to provide a visually appealing image that maintains the essential features of its noisy equivalent. Magnetic Resonance (MR) images are acquired with degradations, and a common deterioration is Rician noise, which arises from variations in temperature or technical faults. Random noise reduces the clarity of images and raises the risk of incorrect diagnosis because it potentially conceals critical anatomical features and important diagnostic observations. Denoising optimizes the visibility of subtle lesions by minimizing noise and increasing diagnostic precision and sensitivity. Various existing denoised methods fail to attenuate the noise properly, leading to blurring or removing fine details from the processed images. Thus, this study proposes an Adapted Fast Gradient Projection (AFGP) algorithm for MR image denoising. The proposed algorithm can automatically compute the regularization parameter for each MR image via the local image information. Moreover, a detail-emphasized phase is applied at each iteration to maintain the structure and delicate features. The performance of the proposed AFGP algorithm is assessed with a dataset of real noisy images, compared with various denoising algorithms, and the results are evaluated using three sophisticated accuracy methods in addition to runtime. Ultimately, the proposed approach yielded satisfactory outcomes, surpassing all comparable techniques with relatively fast runtimes.

Keywords—Magnetic Resonance Images (MRI), denoising, Rician

Cite: Manar A. Al-Abaji and Zohair Al-Ameen, "Adapted Fast Gradient Projection Algorithm for Magnetic Resonance Image Denoising," Journal of Image and Graphics, Vol. 13, No. 2, pp. 140-150, 2025.

Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC-BY-4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.