用移除代替模糊，实现2D/3D视觉任务的隐私保护
Beyond Anonymization: Object Scrubbing for Privacy-Preserving 2D and 3D Vision Tasks
- 通过实例分割+生成修复移除敏感物体，保持场景完整
- 2D检测任务保留87.5%基准精度，优于图像删除的74.2%
- 3D重建中仅损失1.66dB PSNR，感知质量反而提升
我们提出ROAR（鲁棒物体移除与重标注）框架，一种可扩展的隐私保护数据集混淆方法，通过彻底移除敏感物体而非修改它们来实现隐私保护。该方法结合实例分割与生成式修复技术，在消除可识别实体的同时保持场景完整性。在基于COCO的2D目标检测任务中，ROAR达到基线平均精度（AP）的87.5%，而图像删除方法仅达74.2%，凸显了物体移除在维持数据集效用上的优势；小物体因遮挡和细粒度细节丢失导致性能下降更显著。在基于NeRF的3D重建任务中，本方法最大仅造成1.66 dB的PSNR损失，同时保持并提升了SSIM与LPIPS指标，表明其具备优越的感知质量。研究结果确立了物体移除作为有效隐私框架，可在极小性能代价下实现强隐私保障。同时揭示了生成修复、遮挡鲁棒分割与任务特定擦除等关键挑战，为未来隐私保护视觉系统的发展奠定基础。
原文摘要 · Abstract (English)
We introduce ROAR (Robust Object Removal and Re-annotation), a scalable framework for privacy-preserving dataset obfuscation that eliminates sensitive objects instead of modifying them. Our method integrates instance segmentation with generative inpainting to remove identifiable entities while preserving scene integrity. Extensive evaluations on 2D COCO-based object detection show that ROAR achieves 87.5% of the baseline detection average precision (AP), whereas image dropping achieves only 74.2% of the baseline AP, highlighting the advantage of scrubbing in preserving dataset utility. The degradation is even more severe for small objects due to occlusion and loss of fine-grained details. Furthermore, in NeRF-based 3D reconstruction, our method incurs a PSNR loss of at most 1.66 dB while maintaining SSIM and improving LPIPS, demonstrating superior perceptual quality. Our findings establish object removal as an effective privacy framework, achieving strong privacy guarantees with minimal performance trade-offs. The results highlight key challenges in generative inpainting, occlusion-robust segmentation, and task-specific scrubbing, setting the foundation for future advancements in privacy-preserving vision systems.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。