Image restoration based on the fast marching method and block based sampling

摘要

In this paper, we propose an efficient image inpainting algorithm by introducing important aspects and improvements corresponding to the filling order of the pixels in the target region and texture synthesis in a dynamic searching range. The algorithm consists of two parts. The first part decides the filling order of the pixels in the target regions based on the high accuracy fast marching method. The second part of the algorithm implicitly assumes a Markov random field model for textured image regions and computes blocks of texture using an efficient search process and the SSD (Sum of Squared Differences) measure. The algorithm is straightforward to implement and restores the target regions with visually plausible quality that is at par or better than several existing methods, with a lower execution cost.