Trustworthy Image Super-Resolution via Generative Pseudoinverse
This addresses the problem of reliable image restoration for applications requiring high-fidelity upscaling, representing a novel method rather than an incremental improvement.
The paper tackles trustworthy image super-resolution by developing generative models that respect the degradation process and achieve asymptotic consistency with low-resolution measurements, outperforming existing methods by a large margin.
We consider the problem of trustworthy image restoration, taking the form of a constrained optimization over the prior density. To this end, we develop generative models for the task of image super-resolution that respect the degradation process and that can be made asymptotically consistent with the low-resolution measurements, outperforming existing methods by a large margin in that respect.