---
格式版本: 2
标题: "Learning-Based Reconstruction Attacks on Coordinate-Obfuscated Point Clouds"
原文链接: "https://arxiv.org/abs/2609.02568"
发布日期: "2026-09-02"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Wed, 2 Sep 2026 13:17:38 UTC (547 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-09-04T01:39:25+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-09-04T01:39:03+08:00"
入库时间: "2026-09-03T17:39:25.102Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=Immersion&searchtype=all"
匹配关键词:
  - "Immersion"
  - "delivery"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 2
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "该论文研究点云加密与重建攻击，与超节点/AI Rack/机柜级AI基础设施、供电散热互连等完全无关，仅因搜索词Immersion误命中。"
AI质检模型: "zj-deepseek-v4-flash"
AI质检时间: "2026-09-04T01:40:03+08:00"
AI主题相关性: 0
AI来源权威性: 1
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 1
AI摘要: "该研究评估了选择性点云坐标加密方案对机器学习重建攻击的鲁棒性，使用PointNet和随机森林模型测试两种加密粒度。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-04T00:17:35.808Z"
采集批次: "2026年9月3日22点43分34秒"
采集批次ID: "20260903-224334-406"
去重键: "https://arxiv.org/abs/2609.02568"
---

## Computer Science > Cryptography and Security

## Title:Learning-Based Reconstruction Attacks on Coordinate-Obfuscated Point Clouds

[View PDF](https://arxiv.org/pdf/2609.02568) [HTML (experimental)](https://arxiv.org/html/2609.02568v1)

> Abstract:Volumetric video based on point cloud representations enables immersive virtual and augmented reality applications but introduces significant challenges for efficient and secure content delivery. Prior work proposed a selective coordinate encryption framework for point clouds that encrypts only a subset of coordinates, reducing computational costs while visually degrading unauthorized content. However, it remains unclear whether the remaining unencrypted information is sufficient to enable content reconstruction.  
> In this paper, we evaluate the robustness of selective coordinate encryption against machine learning-based reconstruction attacks. We consider an attacker with access to selectively encrypted point clouds attempting to recover encrypted coordinates without decryption by exploiting spatial and geometric correlations in the unencrypted data. We evaluate PointNet and Random Forest models under two encryption granularities: \\texttt{X}, where all $X$ coordinates are encrypted, and \\texttt{2X}, where every second $X$ coordinate is encrypted.  
> Our results show that reconstructing fully encrypted $X$ coordinates remains challenging, whereas the \\texttt{2X} scheme leaks sufficient information through neighboring coordinates to enable accurate reconstruction. These findings demonstrate that the security of selective coordinate encryption depends strongly on encryption granularity.

| Comments: |  |
| --- | --- |
| Subjects: | Cryptography and Security (cs.CR); Machine Learning (cs.LG) |
| Cite as: | [arXiv:2609.02568](https://arxiv.org/abs/2609.02568) \[cs.CR\] |
|  | (or [arXiv:2609.02568v1](https://arxiv.org/abs/2609.02568v1) \[cs.CR\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2609.02568](https://doi.org/10.48550/arXiv.2609.02568) |
| Related DOI: | [https://doi.org/10.1145/3842203.3844590](https://doi.org/10.1145/3842203.3844590) |

## Submission history

From: Mohammad Waquas Usmani \[[view email](https://arxiv.org/show-email/1029f701/2609.02568)\]  
**\[v1\]** Wed, 2 Sep 2026 13:17:38 UTC (547 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2609.02568) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))