mlsanityData to 14 Oct 2026

A Unified Spatial Alignment Framework for Highly Transferable Transformation-Based Attacks on Spatially Structured Tasks

arXiv 2603.25230 ↗First posted 26 Mar 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Transformation-based adversarial attacks (TAAs) demonstrate strong transferability when deceiving classification models. However, existing TAAs often perform unsatisfactorily or even fail when applied to structured tasks such as semantic segmentation and object detection. Encouragingly, recent studies that categorize transformations into non-spatial and spatial transformations inspire us to address this challenge. We find that for non-structured tasks, labels are spatially non-structured, and thus TAAs are not required to adjust labels when applying spatial transformations. In contrast, for structured tasks, labels are spatially structured, and failing to transform labels synchronously with inputs can cause spatial misalignment and yield erroneous gradients. To address these issues, we propose a novel unified Spatial Alignment Framework (SAF) for highly transferable TAAs on spatially structured tasks, where the TAAs spatially transform labels synchronously with the input using the proposed Spatial Alignment (SA) algorithm. Extensive experiments demonstrate the crucial role of our SAF for TAAs on structured tasks. Specifically, in non-targeted attacks, our SAF degrades the average mIoU on Cityscapes from 24.50 to 11.34, and on Kvasir-SEG from 49.91 to 31.80, while reducing the average mAP of COCO from 17.89 to 5.25.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 42 of its 57 references that are in the map. First posted on arXiv 26 Mar 2026.

Builds on

42 of 57 references.

References · 42

  1. 2014Microsoft COCO: Common Objects in Contextcited by 17,608
  2. 2015Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networkscited by 9,639
  3. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  4. 2016Feature Pyramid Networks for Object Detectioncited by 5,036
  5. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,444
  6. 2018Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentationcited by 3,223
  7. 2015SSD: Single Shot MultiBox Detectorcited by 3,130
  8. 2021SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformerscited by 2,940
  9. 2018YOLOv3: An Incremental Improvementcited by 2,659
  10. 2016Pyramid Scene Parsing Networkcited by 2,479
  11. 2017Rethinking Atrous Convolution for Semantic Image Segmentationcited by 2,161
  12. 2019FCOS: Fully Convolutional One-Stage Object Detectioncited by 1,577
30 more
  1. 2018UNet++: A Nested U-Net Architecture for Medical Image Segmentationcited by 1,525
  2. 2018Attention U-Net: Learning Where to Look for the Pancreascited by 1,435
  3. 2019Deep High-Resolution Representation Learning for Visual Recognitioncited by 1,420
  4. 2018Unified Perceptual Parsing for Scene Understandingcited by 1,271
  5. 2018Improving Transferability of Adversarial Examples with Input Diversitycited by 599
  6. 2019Kvasir-SEG: A Segmented Polyp Datasetcited by 589
  7. 2017Countering Adversarial Images using Input Transformationscited by 520
  8. 2024YOLOv10: Real-Time End-to-End Object Detectioncited by 515
  9. 2019Evading Defenses to Transferable Adversarial Examples by Translation-Invariant Attackscited by 442
  10. 2019Nesterov Accelerated Gradient and Scale Invariance for Adversarial Attackscited by 355
  11. 2025YOLOv12: Attention-Centric Real-Time Object Detectorscited by 305
  12. 2017Adversarial Examples for Semantic Segmentation and Object Detectioncited by 275
  13. 2019ResUNet-a: a deep learning framework for semantic segmentation of remotely sensed datacited by 261
  14. 2021Admix: Enhancing the Transferability of Adversarial Attackscited by 156
  15. 2022Frequency Domain Model Augmentation for Adversarial Attackcited by 130
  16. 2017Keeping the Bad Guys Out: Protecting and Vaccinating Deep Learning with JPEG Compressioncited by 87
  17. 2020Uncertainty-aware multi-view co-training for semi-supervised medical image segmentation and domain adaptationcited by 81
  18. 2022SegPGD: An Effective and Efficient Adversarial Attack for Evaluating and Boosting Segmentation Robustnesscited by 74
  19. 2023Boosting Adversarial Transferability by Block Shuffle and Rotationcited by 64
  20. 2023Structure Invariant Transformation for better Adversarial Transferabilitycited by 62
  21. 2021Improving the Transferability of Adversarial Examples with Resized-Diverse-Inputs, Diversity-Ensemble and Region Fittingcited by 37
  22. 2023CosPGD: an efficient white-box adversarial attack for pixel-wise prediction taskscited by 27
  23. 2018Robust Adversarial Perturbation on Deep Proposal-based Modelscited by 26
  24. 2024Learning to Transform Dynamically for Better Adversarial Transferabilitycited by 23
  25. 2021Adaptive Image Transformations for Transfer-based Adversarial Attackcited by 22
  26. 2021Context-Aware Transfer Attacks for Object Detectioncited by 18
  27. 2024Boosting Adversarial Transferability across Model Genus by Deformation-Constrained Warpingcited by 18
  28. 2022Zero-Query Transfer Attacks on Context-Aware Object Detectorscited by 17
  29. 2023Improving the Transferability of Adversarial Examples with Arbitrary Style Transfercited by 13
  30. 2023TranSegPGD: Improving Transferability of Adversarial Examples on Semantic Segmentationcited by 10

Built on by

Not cited yet.