Primal-Dual Method for Image Denoising with Variable Exponent Sobolev Spaces
Hamdi Braiek · PRSM · 2025
In this work, we present a novel approach for image denoising based on a primal-dual algorithm within variable exponent Sobolev spaces $W^{1,p(x)}$. This model adapts to local image features, effectively balancing noise reduction and detail preservation. The performance of the proposed method is evaluated on grayscale and color images degraded by Gaussian noise and is compared to the total variation and Beltrami models. Results show that our approach outperforms the alternatives in terms of noise suppression, detail preservation, and convergence speed.