SR-Prominence: A Crowdsourced Protocol and Dataset Suite for Perceptually-Weighted Super-Resolution Artifact Evaluation

Ivan Molodetskikh, Kirill Malyshev, Mark Mirgaleev, Evgeney Bogatyrev, Nikita Zagainov, Dmitriy Vatolin · 2026

Read SR-Prominence: A Crowdsourced Protocol and Dataset Suite for Perceptually-Weighted Super-Resolution Artifact Evaluation by Ivan Molodetskikh, Kirill Malyshev, Mark Mirgaleev, Evgeney Bogatyrev, Nikita Zagainov, Dmitriy Vatolin (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-14847.txt.

Description

Molodetskikh and team propose artifact prominence for super-resolution scoring, defined as the fraction of viewers judging a marked region to contain a visible artifact. Existing detectors treat all flaws equally, though visual impact varies widely. A crowdsourced protocol builds SR-Prominence, with nearly four thousand flaw masks from several sources, annotated with scores. Re-annotating one source reveals about forty-eight percent of its in-lab flaws go unnoticed by most viewers. Classical full-reference metrics give surprisingly strong local signals.

Catalog record

CREATORIvan Molodetskikh, Kirill Malyshev, Mark Mirgaleev, Evgeney Bogatyrev, Nikita Zagainov, Dmitriy Vatolin
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53281
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.14847

Browse the full library · About the Tower of Babel