Kaleidoscopic Background Attack: Disrupting Pose Estimation with Multi-Fold Radial Symmetry Textures
Not on the map. Why?
Abstract (arXiv)
Camera pose estimation is a fundamental computer vision task that is essential for applications like visual localization and multi-view stereo reconstruction. In the object-centric scenarios with sparse inputs, the accuracy of pose estimation can be significantly influenced by background textures that occupy major portions of the images across different viewpoints. In light of this, we introduce the Kaleidoscopic Background Attack (KBA), which uses identical segments to form discs with multi-fold radial symmetry. These discs maintain high similarity across different viewpoints, enabling effective attacks on pose estimation models even with natural texture segments. Additionally, a projected orientation consistency loss is proposed to optimize the kaleidoscopic segments, leading to significant enhancement in the attack effectiveness. Experimental results show that optimized adversarial kaleidoscopic backgrounds can effectively attack various camera pose estimation models.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 34 of its 57 references that are in the map. First posted on arXiv 14 Jul 2025.
Builds on
34 of 57 references.
References · 34
- 2014Explaining and Harnessing Adversarial Examplescited by 7,364
- 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
- 2013Intriguing properties of neural networkscited by 5,155
- 2016Towards Evaluating the Robustness of Neural Networkscited by 2,998
- 2016Adversarial examples in the physical worldcited by 1,769
- 2023DUSt3R: Geometric 3D Vision Made Easycited by 1,567
- 2019SuperGlue: Learning Feature Matching with Graph Neural Networkscited by 1,443
- 2017SuperPoint: Self-Supervised Interest Point Detection and Descriptioncited by 1,320
- 2024Grounding Image Matching in 3D with MASt3Rcited by 983
- 2020Accelerating 3D Deep Learning with PyTorch3Dcited by 784
- 2021Common Objects in 3D: Large-Scale Learning and Evaluation of Real-life 3D Category Reconstructioncited by 652
- 2017Adversarial Patchcited by 517
22 more
- 2023OmniObject3D: Large-Vocabulary 3D Object Dataset for Realistic Perception, Reconstruction and Generationcited by 326
- 2023PoseDiffusion: Solving Pose Estimation via Diffusion-aided Bundle Adjustmentcited by 147
- 2024Cameras as Rays: Pose Estimation via Ray Diffusioncited by 131
- 2018LaVAN: Localized and Visible Adversarial Noisecited by 113
- 20243D Geometry-aware Deformable Gaussian Splatting for Dynamic View Synthesiscited by 109
- 2022RelPose: Predicting Probabilistic Relative Rotation for Single Objects in the Wildcited by 102
- 2022Adversarial Texture for Fooling Person Detectors in the Physical Worldcited by 96
- 2023RelPose++: Recovering 6D Poses from Sparse-view Observationscited by 87
- 2023Neural Parametric Gaussians for Monocular Non-Rigid Object Reconstructioncited by 61
- 2022T-SEA: Transfer-based Self-Ensemble Attack on Object Detectioncited by 59
- 2022SparsePose: Sparse-View Camera Pose Regression and Refinementcited by 53
- 2022Few-View Object Reconstruction with Unknown Categories and Camera Posescited by 53
- 2022Simultaneously Optimizing Perturbations and Positions for Black-box Adversarial Patch Attackscited by 42
- 2018Rethinking Feature Distribution for Loss Functions in Image Classificationcited by 39
- 2023Physically Adversarial Infrared Patches with Learnable Shapes and Locationscited by 39
- 2020Physical Adversarial Attack on Vehicle Detector in the Carla Simulatorcited by 38
- 2022The 8-Point Algorithm as an Inductive Bias for Relative Pose Prediction by ViTscited by 38
- 2021Robust Feature-Level Adversaries are Interpretability Toolscited by 27
- 2023Carve3D: Improving Multi-view Reconstruction Consistency for Diffusion Models with RL Finetuningcited by 22
- 2024Learning to Produce Semi-dense Correspondences for Visual Localizationcited by 10
- 2020Shaping Deep Feature Space towards Gaussian Mixture for Visual Classificationcited by 5
- 2024Generalizable Novel-View Synthesis using a Stereo Cameracited by 5
Built on by
Citing papers per year: 2026: 1.
1 paper.