PatchPoison: Poisoning Multi-View Datasets to Degrade 3D Reconstruction

2026-04-14

AI Summary

This work proposes a plug-and-play data poisoning method to prevent unauthorized high-fidelity 3D reconstruction from publicly available multi-view images. The approach embeds imperceptible, high-frequency adversarial checkerboard perturbations along image edges, specifically designed to disrupt feature matching in Structure-from-Motion (SfM) pipelines. This interference induces erroneous camera pose estimation, thereby compromising the geometric consistency required for subsequent 3D Gaussian Splatting reconstruction. Notably, the method requires no modification to the reconstruction pipeline and is compatible with mainstream SfM systems such as COLMAP. Experiments on the NeRF-Synthetic benchmark demonstrate that inserting a mere 12×12-pixel perturbation patch increases the LPIPS reconstruction error by 6.8×, substantially degrading reconstruction quality while remaining nearly invisible to human observers.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

3D Gaussian Splatting (3DGS) has recently enabled highly photorealistic 3D reconstruction from casually captured multi-view images. However, this accessibility raises a privacy concern: publicly available images or videos can be exploited to reconstruct detailed 3D models of scenes or objects without the owner's consent. We present PatchPoison, a lightweight dataset-poisoning method that prevents unauthorized 3D reconstruction. Unlike global perturbations, PatchPoison injects a small high-frequency adversarial patch, a structured checkerboard, into the periphery of each image in a multi-view dataset. The patch is designed to corrupt the feature-matching stage of Structure-from-Motion (SfM) pipelines such as COLMAP by introducing spurious correspondences that systematically misalign estimated camera poses. Consequently, downstream 3DGS optimization diverges from the correct scene geometry. On the NeRF-Synthetic benchmark, inserting a 12 X 12 pixel patch increases reconstruction error by 6.8x in LPIPS, while the poisoned images remain unobtrusive to human viewers. PatchPoison requires no pipeline modifications, offering a practical, "drop-in" preprocessing step for content creators to protect their multi-view data.

Problem

• 3D reconstruction
• privacy
• dataset poisoning
• multi-view images
• 3D Gaussian Splatting

Innovation

• adversarial patch
• dataset poisoning
• 3D reconstruction
• Structure-from-Motion
• 3D Gaussian Splatting