Image restoration, which is an inverse problem such as compressed
sensing and super-resolution, requires priors (e.g. Total Variation
minimization) for regularization.
In this paper, we propose a new image restoration method using a lossy
image codec.
With a prior which is implicitly defined by an off-the-shelf image
codec (e.g HEIF), the proposed method can recover targets efficiently.
In our experiment, the proposed method outperformed hand-crafted prior-based methods.