This paper introduces an iterative regularization method for image restoration based on Bregman distances, with a focus on total variation (TV) regularization. The method improves upon the classical Rudin-Osher-Fatemi (ROF) model by iteratively refining the solution using Bregman distances, leading to better denoising and deblurring results. The key idea is to use the Bregman distance associated with the TV functional to iteratively minimize a modified ROF model, which allows for a more accurate reconstruction of the true image. The method is shown to converge monotonically to the true noise-free image, and the stopping criterion is based on the discrepancy principle, ensuring that the residual error is comparable to the noise level. The algorithm is analyzed for its convergence properties, and it is shown that the sequence of iterates converges weakly-* to the true solution in the space of functions of bounded variation. The method is also applied to other inverse problems, including deblurring and image inpainting, and it is demonstrated that the iterative regularization procedure can be generalized to other regularization models and constraints. The results show that the proposed method provides significant improvements over standard models in denoising and deblurring tasks, and it is shown to be effective in separating cartoon-like structures from textures in images. The method is also analyzed for its behavior under noisy data, and it is shown that the stopping criterion based on the discrepancy principle ensures that the algorithm converges to the true solution as the noise level approaches zero. The paper concludes with a discussion of the theoretical and practical implications of the proposed method for image restoration and inverse problems.This paper introduces an iterative regularization method for image restoration based on Bregman distances, with a focus on total variation (TV) regularization. The method improves upon the classical Rudin-Osher-Fatemi (ROF) model by iteratively refining the solution using Bregman distances, leading to better denoising and deblurring results. The key idea is to use the Bregman distance associated with the TV functional to iteratively minimize a modified ROF model, which allows for a more accurate reconstruction of the true image. The method is shown to converge monotonically to the true noise-free image, and the stopping criterion is based on the discrepancy principle, ensuring that the residual error is comparable to the noise level. The algorithm is analyzed for its convergence properties, and it is shown that the sequence of iterates converges weakly-* to the true solution in the space of functions of bounded variation. The method is also applied to other inverse problems, including deblurring and image inpainting, and it is demonstrated that the iterative regularization procedure can be generalized to other regularization models and constraints. The results show that the proposed method provides significant improvements over standard models in denoising and deblurring tasks, and it is shown to be effective in separating cartoon-like structures from textures in images. The method is also analyzed for its behavior under noisy data, and it is shown that the stopping criterion based on the discrepancy principle ensures that the algorithm converges to the true solution as the noise level approaches zero. The paper concludes with a discussion of the theoretical and practical implications of the proposed method for image restoration and inverse problems.