Spectral Robustness Mixer: Cross-Scale Neck for Robust No-Reference Image Quality Assessment
Technologies · 28.02.2026 · 14 (3) · 145
DOI: 10.3390/technologies14030145 ↗Kurz gefasst
Fachartikel in der Zeitschrift Technologies, online erschienen am 28.02.2026. Verfasst von Bader Rasheed und 2 weiteren Autor:innen. Die Arbeit nennt 49 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministry of Economic Development of the Russian Federation.
Abstract (Originalsprache)
No-reference image quality assessment (NR-IQA) models achieve high correlation with human mean opinion scores (MOS) on clean benchmarks, yet recent work shows they can be highly vulnerable to small adversarial perturbations that severely degrade ranking consistency, including in black-box settings. We introduce the Spectral Robustness Mixer (SRM), a lightweight neck inserted between an NR-IQA backbone and regression head, designed to reduce adversarial sensitivity without changing the dataset, label format, or target metric. SRM couples (i) deep-to-shallow cross-scale fusion via a Nyström low-rank attention surrogate, (ii) ridge-conditioned landmark kernels with ridge regularization, solved via numerically stable small-matrix factorization (SVD/LU) to improve conditioning, and (iii) variance-aware entropy-regularized fusion gates with a bounded gain cap to limit gradient amplification. We evaluate SRM on TID2013 and KonIQ-10k under a white-box l∞/l2 attack ensemble that includes per-image regression objectives and a correlation-aware pairwise inversion objective (a ranking-inspired surrogate for correlation inversion), with expectation-over-transformation (EOT) and anti-gradient masking checks. At ϵ=4/255 (l∞), SRM improves worst-case robust Spearman’s rank-order correlation coefficient (SROCC; defined as the minimum over our fixed attack ensemble) by an absolute 0.06–0.08 SROCC points (i.e., correlation-coefficient units, not percentage gain) across datasets/backbones, while keeping clean SROCC within 0.00–0.01 of the baseline. We observe similar trends for Pearson linear correlation coefficient (PLCC).
Förderung
- Ministry of Economic Development of the Russian Federation
Zitieren
APA
Rasheed, B., Antsiferova, A., & Vatolin, D. (2026). Spectral Robustness Mixer: Cross-Scale Neck for Robust No-Reference Image Quality Assessment. Technologies, 14(3), 145. https://doi.org/10.3390/technologies14030145
Vancouver
Rasheed B, Antsiferova A, Vatolin D. Spectral Robustness Mixer: Cross-Scale Neck for Robust No-Reference Image Quality Assessment. Technologies. 2026;14(3):145. doi:10.3390/technologies14030145
BibTeX
@article{Rasheed2026,
author = {Rasheed, Bader and Antsiferova, Anastasia and Vatolin, Dmitriy},
title = {Spectral Robustness Mixer: Cross-Scale Neck for Robust No-Reference Image Quality Assessment},
journal = {Technologies},
year = {2026},
volume = {14},
number = {3},
pages = {145},
doi = {10.3390/technologies14030145}
}
EUVIMED