English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In this work, we utilize the high-fidelity generation abilities of diffusionmodels to solve blind JPEG restoration at high compression levels. We proposean elegant modification of the forward stochastic differential equation ofdiffusion models to adapt them to this restoration task and name our methodDriftRec. Comparing DriftRec against an $L_2$ regression baseline with the samenetwork architecture and two state-of-the-art techniques for JPEG restoration,we show that our approach can escape the tendency of other methods to generateblurry images, and recovers the distribution of clean images significantly morefaithfully. For this, only a dataset of clean/corrupted image pairs and noknowledge about the corruption operation is required, enabling widerapplicability to other restoration tasks. In contrast to other conditional andunconditional diffusion models, we utilize the idea that the distributions ofclean and corrupted images are much closer to each other than each is to theusual Gaussian prior of the reverse process in diffusion models. Our approachtherefore requires only low levels of added noise, and needs comparatively fewsampling steps even without further optimizations. We show that DriftRecnaturally generalizes to realistic and difficult scenarios such as unaligneddouble JPEG compression and blind restoration of JPEGs found online, withouthaving encountered such examples during training.

DriftRec: Adapting diffusion models to blind JPEG restoration