This paper introduces a novel algorithm for digital image inpainting, aiming to replicate the techniques used by professional restorators. The algorithm automatically fills-in selected regions with information from surrounding areas, without requiring user intervention beyond marking the regions to be inpainted. The fill-in process is designed to maintain the continuity of isophote lines (lines of equal gray values) at the boundaries of the regions being inpainted. The algorithm is robust to different backgrounds and does not impose restrictions on the topology of the regions. It can be applied to various tasks such as restoring old photographs, removing superimposed text, and removing objects from images. The paper discusses the limitations of existing techniques, such as those for film restoration and texture synthesis, and highlights the advantages of the proposed method. The implementation details, including the discrete scheme and numerical stability considerations, are provided, along with examples demonstrating the effectiveness of the algorithm in restoring images with complex structures and textures.This paper introduces a novel algorithm for digital image inpainting, aiming to replicate the techniques used by professional restorators. The algorithm automatically fills-in selected regions with information from surrounding areas, without requiring user intervention beyond marking the regions to be inpainted. The fill-in process is designed to maintain the continuity of isophote lines (lines of equal gray values) at the boundaries of the regions being inpainted. The algorithm is robust to different backgrounds and does not impose restrictions on the topology of the regions. It can be applied to various tasks such as restoring old photographs, removing superimposed text, and removing objects from images. The paper discusses the limitations of existing techniques, such as those for film restoration and texture synthesis, and highlights the advantages of the proposed method. The implementation details, including the discrete scheme and numerical stability considerations, are provided, along with examples demonstrating the effectiveness of the algorithm in restoring images with complex structures and textures.