PrinGS: Privacy Preservation in Gaussian Splatting
Computer Science > Computer Vision and Pattern Recognition arXiv:2610.36168 (cs) [Submitted on 26 Oct 2026] Title: PrinGS: Privacy Preservation in Gaussian Splatting Authors: Gaurav Behera, Abhinav Raundhal, Archisha Panda, Saumya Balina, Gargi Shroff, Ponnurangam Kumaraguru.
计算机科学 > 计算机视觉与模式识别 arXiv:2610.36168 (cs) [提交于 2026 年 10 月 26 日] 标题：PrinGS：高斯溅射中的隐私保护 作者：Gaurav Behera, Abhinav Raundhal, Archisha Panda, Saumya Balina, Gargi Shroff, Ponnurangam Kumaraguru。
Abstract: 3D Gaussian Splatting (3DGS) has rapidly become the preferred representation for high-fidelity novel view synthesis and 3D scene reconstruction. However, the learned representation encodes sensitive visual information including human faces, license plates, documents, and computer screens, raising significant privacy concerns when trained splats are shared or deployed.
摘要：3D 高斯溅射 (3DGS) 已迅速成为高保真新视角合成和 3D 场景重建的首选表示方法。然而，学习到的表示编码了敏感的视觉信息，包括人脸、车牌、文档和电脑屏幕，当训练好的溅射模型被共享或部署时，这引发了重大的隐私担忧。
In this work, we address the underexplored problem of privacy preservation in Gaussian Splatting by proposing a representation-level anonymization framework that directly anonymizes a trained splat. Given a trained splat and a user-provided reference view identifying sensitive content, our method localizes the corresponding regions in the underlying 3D representation and directly anonymizes them, ensuring that the sensitive information is absent from all subsequently rendered novel views while preserving the overall scene geometry and visual fidelity.
在这项工作中，我们通过提出一种表示层面的匿名化框架来解决高斯溅射中隐私保护这一未被充分探索的问题，该框架可以直接对训练好的溅射模型进行匿名化处理。给定一个训练好的溅射模型和用户提供的识别敏感内容参考视图，我们的方法可以定位底层 3D 表示中的相应区域并直接对其进行匿名化，确保敏感信息在随后渲染的所有新视角中均不存在，同时保持整体场景的几何结构和视觉保真度。
We achieve this post-training anonymization in less than 1 second. Furthermore, we extend our framework with a privacy-by-design training strategy that prevents sensitive content from being embedded into the representation during optimization, enabling the generation of privacy-preserving splats directly from training images.
我们在不到 1 秒的时间内实现了这种训练后的匿名化。此外，我们通过一种“隐私设计”训练策略扩展了我们的框架，该策略可防止敏感内容在优化过程中被嵌入到表示中，从而能够直接从训练图像生成具有隐私保护功能的溅射模型。
Our method achieves a remarkably low Face Similarity Measure (FSM) of 9.79% at a retained Face Detection Rate (FDR) of 39.47% and makes sensitive text completely unreadable with an OCR Character Error Rate of 1.00. Our work aims to serve as the primary stepping stone towards privacy-aware 3D representations and motivates further research along these directions.
我们的方法在保持 39.47% 的人脸检测率 (FDR) 的情况下，实现了 9.79% 的极低人脸相似度度量 (FSM)，并使敏感文本完全无法读取，OCR 字符错误率达到 1.00。我们的工作旨在成为迈向隐私感知 3D 表示的首要基石，并推动这些方向的进一步研究。