Fspgd: rethinking black-box attacks on semantic segmentation
Machine Vision and Applications · 2026-11 · 37 (6)
DOI: 10.1007/s00138-026-01919-5 ↗Kurz gefasst
Fachartikel in der Zeitschrift Machine Vision and Applications, online erschienen am 04.09.2026. Verfasst von Eun-Sol Park und 2 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 48 Literaturangaben.
Abstract (Originalsprache)
Black-box adversarial attacks on semantic segmentation remain a challenging problem, particularly in the black-box transfer attack setting where perturbations crafted on a surrogate model are expected to mislead unseen target models. Existing methods typically operate only on output logits and thus fail to account for the spatial structure and class-wise feature relationships that are crucial for dense prediction. To address this limitation, we propose Feature Similarity Projected Gradient Descent (FSPGD), a feature-space black-box attack that explicitly disrupts intermediate representations. FSPGD employs a dual loss design: an external loss that enforces discrepancy between clean and adversarial features to weaken cross-model alignment, and an internal loss that reduces feature consistency among spatially separated instances of the same class. Comprehensive experiments on Pascal VOC 2012 and Cityscapes across both CNN-based and Transformer-based backbones demonstrate that FSPGD achieves state-of-the-art transferability, consistently outperforming conventional logit-level methods as well as recent segmentation-specific baselines such as SegPGD, CosPGD, and RP-PGD. Moreover, adversarial training with FSPGD examples enhances robustness against unseen attacks across multiple architectures, further validating the effectiveness of our design. These findings establish FSPGD as a principled and practical framework for advancing black-box adversarial attacks in semantic segmentation. Code is available at https://github.com/KU-AIVS/FSPGD.
Zitieren
APA
Park, E., Park, M., & Shin, Y. (2026). Fspgd: rethinking black-box attacks on semantic segmentation. Machine Vision and Applications, 37(6). https://doi.org/10.1007/s00138-026-01919-5
Vancouver
Park E, Park M, Shin Y. Fspgd: rethinking black-box attacks on semantic segmentation. Machine Vision and Applications. 2026;37(6). doi:10.1007/s00138-026-01919-5
BibTeX
@article{Park2026,
author = {Park, Eun-Sol and Park, Miso and Shin, Yong-Goo},
title = {Fspgd: rethinking black-box attacks on semantic segmentation},
journal = {Machine Vision and Applications},
year = {2026},
volume = {37},
number = {6},
doi = {10.1007/s00138-026-01919-5}
}
EUVIMED