Rank Minimization-Based Fast Image Completion

We propose an image completion algorithm using a rank minimization framework. Based on the lowrank property of an image, we formulate image completion as a low-rank matrix completion problem. Then, we solve the optimization problem efficiently using the augmented Lagrange multiplier (ALM) method. Experiments show that the proposed algorithm provides comparable or even higher image qualities than state-of-the-art approaches, while demanding significantly lower computational resources.