From Posterior Sampling to Meaningful Diversity in Image Restoration

arXiv:2310.16047v2 Announce Type: replace-cross
Abstract: Image restoration problems are typically ill-posed in the sense that each degraded image can be restored in infinitely many valid ways. To accommodate this, many works generate a diverse set of outputs by attempting to randomly sample from the posterior distribution of natural images given the degraded input. Here we argue that this strategy is commonly of limited practical value because of the heavy tail of the posterior distribution. Consider for example inpainting a missing …

arxiv cs.cv cs.lg diversity eess.iv image image restoration posterior sampling type