A Three-term PRP-DL Method Modification with Application in Image Denoising Problem

Abstract

Image denoising poses a critical challenge due to the impact of noise on image quality and the need to preserve essential details. This study introduces a hybrid Polak-Ribiére-Polyak (PRP)-Dai-Liao (DL) conjugate gradient method with a modified scalar to improve the performance of denoising algorithms on large-scale images. The proposed method involves modifying the scalar in the PRP-DL conjugate gradient method, thereby enhancing algorithmic efficiency, especially in handling large-scale problems. Convergence analysis under the standard Wolfe-Powell line search is established, and numerical results demonstrate that the proposed method is more efficient and robust than existing conjugate gradient methods. The application of the method to image denoising with various noise levels and window sizes confirms its capability to effectively remove noise while preserving image details. Overall, this modified conjugate gradient method shows promise for practical applications in image denoising problem.