# PoInit-of-View: Poisoning Initialization of Views Transfers Across Multiple 3D Reconstruction Systems

arXiv: https://arxiv.org/abs/2604.16540 · HTML: https://mlsanity.com/paper/2604.16540/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 42 of its 57 references that are in the map.
- First posted on arXiv 17 Apr 2026.

## Builds on (most cited first, 30 of 42)

- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [3D Gaussian Splatting for Real-Time Radiance Field Rendering](https://mlsanity.com/paper/2308.04079/) (2023)
- [Instant Neural Graphics Primitives with a Multiresolution Hash Encoding](https://mlsanity.com/paper/2201.05989/) (2022)
- [Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields](https://mlsanity.com/paper/2111.12077/) (2021)
- [MVSNet: Depth Inference for Unstructured Multi-view Stereo](https://mlsanity.com/paper/1804.02505/) (2018)
- [Mip-Splatting: Alias-free 3D Gaussian Splatting](https://mlsanity.com/paper/2311.16493/) (2023)
- [BARF: Bundle-Adjusting Neural Radiance Fields](https://mlsanity.com/paper/2104.06405/) (2021)
- [VAD: Vectorized Scene Representation for Efficient Autonomous Driving](https://mlsanity.com/paper/2303.12077/) (2023)
- [Delving into Transferable Adversarial Examples and Black-box Attacks](https://mlsanity.com/paper/1611.02770/) (2016)
- [DeepVoxels: Learning Persistent 3D Feature Embeddings](https://mlsanity.com/paper/1812.01024/) (2018)
- [NeuralRecon: Real-Time Coherent 3D Reconstruction from Monocular Video](https://mlsanity.com/paper/2104.00681/) (2021)
- [DriveAdapter: Breaking the Coupling Barrier of Perception and Planning in End-to-End Autonomous Driving](https://mlsanity.com/paper/2308.00398/) (2023)
- [On Success and Simplicity: A Second Look at Transferable Targeted Attacks](https://mlsanity.com/paper/2012.11207/) (2020)
- [Rethinking Depth Estimation for Multi-View Stereo: A Unified Representation](https://mlsanity.com/paper/2201.01501/) (2022)
- [Wild-GS: Real-Time Novel View Synthesis from Unconstrained Photo Collections](https://mlsanity.com/paper/2406.10373/) (2024)
- [ViewFool: Evaluating the Robustness of Visual Recognition to Adversarial Viewpoints](https://mlsanity.com/paper/2210.03895/) (2022)
- [Physical 3D Adversarial Attacks against Monocular Depth Estimation in Autonomous Driving](https://mlsanity.com/paper/2403.17301/) (2024)
- [NerfBaselines: Consistent and Reproducible Evaluation of Novel View Synthesis Methods](https://mlsanity.com/paper/2406.17345/) (2024)
- [Robotic Visual Instruction](https://mlsanity.com/paper/2505.00693/) (2025)
- [Poison-splat: Computation Cost Attack on 3D Gaussian Splatting](https://mlsanity.com/paper/2410.08190/) (2024)
- [FewViewGS: Gaussian Splatting with Few View Matching and Multi-stage Training](https://mlsanity.com/paper/2411.02229/) (2024)
- [Real-Time Simulated Avatar from Head-Mounted Sensors](https://mlsanity.com/paper/2403.06862/) (2024)
- [VGOS: Voxel Grid Optimization for View Synthesis from Sparse Inputs](https://mlsanity.com/paper/2304.13386/) (2023)
- [Geometry Cloak: Preventing TGS-based 3D Reconstruction from Copyrighted Images](https://mlsanity.com/paper/2410.22705/) (2024)
- [Benchmarking Robustness in Neural Radiance Fields](https://mlsanity.com/paper/2301.04075/) (2023)
- [T2TD: Text-3D Generation Model based on Prior Knowledge Guidance](https://mlsanity.com/paper/2305.15753/) (2023)
- [Cross-Modal and Uncertainty-Aware Agglomeration for Open-Vocabulary 3D Scene Understanding](https://mlsanity.com/paper/2503.16707/) (2025)
- [Structural Multiplane Image: Bridging Neural View Synthesis and 3D Reconstruction](https://mlsanity.com/paper/2303.05937/) (2023)
- [MAGiC-SLAM: Multi-Agent Gaussian Globally Consistent SLAM](https://mlsanity.com/paper/2411.16785/) (2024)
- [Gaussian Splatting Under Attack: Investigating Adversarial Noise in 3D Objects](https://mlsanity.com/paper/2412.02803/) (2024)

## Built on by (most cited first, 1 of 1)

- [Medoid Prototype Alignment for Cross-Plant Unknown Attack Detection in Industrial Control Systems](https://mlsanity.com/paper/2604.25544/) (2026)

## Abstract (arXiv)

Poisoning input views of 3D reconstruction systems has been recently studied. However, we identify that existing studies simply backpropagate adversarial gradients through the 3D reconstruction pipeline as a whole, without uncovering the new vulnerability rooted in specific modules of the 3D reconstruction pipeline. In this paper, we argue that the structure-from-motion (SfM) initialization, as the geometric core of many widely used reconstruction systems, can be targeted to achieve transferable poisoning effects across diverse 3D reconstruction systems. To this end, we propose PoInit-of-View, which optimizes adversarial perturbations to intentionally introduce cross-view gradient inconsistencies at projections of corresponding 3D points. These inconsistencies disrupt keypoint detection and feature matching, thereby corrupting pose estimation and triangulation within SfM, eventually resulting in low-quality rendered views. We also provide a theoretical analysis that connects cross-view inconsistency to correspondence collapse. Experimental results demonstrate the effectiveness of our PoInit-of-View on diverse 3D reconstruction systems and datasets, surpassing the single-view baseline by 25.1% in PSNR and 16.5% in SSIM in black-box transfer settings, such as 3DGS to NeRF.