EUVIMEDEUVIMED entdecken
Crossref

Wrapper-Based Adversarial Input Screening for Deep Image Classifiers Using Feature Squeezing and Logit-Space Inconsistency

Alketa Hyso, Dezdemona Gjylapi

Journal of Innovative Image Processing · 10.07.2026 · 8 (3) · 870-893

DOI: 10.36548/jiip.2026.3.007 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Innovative Image Processing, online erschienen am 10.07.2026. Verfasst von Alketa Hyso und 1 weiteren Autor:innen. Die Arbeit nennt 24 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Adversarial perturbations pose a practical integrity risk to vision-based decision systems by causing image classifiers to misclassify inputs after visually subtle changes. This paper evaluates a wrapper-based adversarial input screening approach that compares a classifier’s output on an original image with its output after benign feature-squeezing transformations. The evaluated transformations include median filtering, non-local means denoising, and bit-depth reduction. Using a frozen-backbone ResNet50 on CIFAR-10, the detector is assessed under untargeted and targeted Fast Gradient Sign Method attacks, followed by stronger Projected Gradient Descent verification. Detection thresholds are calibrated only on clean data using a fixed 5% false positive rate protocol and are validated on held-out clean samples. The results show that softmax-space ℓ1 inconsistency provides moderate, transformation-dependent detection, whereas logit-space ℓ2 inconsistency yields a stronger, more stable screening signal. Median filtering with logit-space ℓ2 achieves near-complete detection at ε = 8/255 and remains the most reliable configuration across the perturbation sensitivity analysis, while non-local means denoising becomes effective mainly for large perturbations. An additional two-stage gate is evaluated for deployment-oriented alert triage; it does not replace the primary detector or override its screening decision, but ranks flagged inputs by risk severity. Further evaluations show that performance decreases on TinyImageNet and that targeted threshold-aware adaptive optimisation can substantially reduce detection recall. The findings support the use of median filtering with logit-space ℓ2 inconsistency as a tool for screening adversarial inputs to image classifiers, but its effectiveness depends on dataset complexity, classifier behaviour, and attack adaptivity.

Zitieren

APA

Hyso, A., & Gjylapi, D. (2026). Wrapper-Based Adversarial Input Screening for Deep Image Classifiers Using Feature Squeezing and Logit-Space Inconsistency. Journal of Innovative Image Processing, 8(3), 870-893. https://doi.org/10.36548/jiip.2026.3.007

Vancouver

Hyso A, Gjylapi D. Wrapper-Based Adversarial Input Screening for Deep Image Classifiers Using Feature Squeezing and Logit-Space Inconsistency. Journal of Innovative Image Processing. 2026;8(3):870-893. doi:10.36548/jiip.2026.3.007

BibTeX
@article{Hyso2026,
  author = {Hyso, Alketa and Gjylapi, Dezdemona},
  title = {Wrapper-Based Adversarial Input Screening for Deep Image Classifiers Using Feature Squeezing and Logit-Space Inconsistency},
  journal = {Journal of Innovative Image Processing},
  year = {2026},
  volume = {8},
  number = {3},
  pages = {870--893},
  doi = {10.36548/jiip.2026.3.007}
}
Kostenlos starten →