将手语姿势与形象分离，实现身份匿名化同时保持动作自然
Pose-Based Sign Language Appearance Transfer
- 基于骨骼姿态估计，迁移不同手语者外观但保留动作内容
- 使手语识别准确率轻微下降，但显著降低身份辨识度
- 适合需要隐私保护的手语视频处理场景
我们提出一种在保持手语内容不变的前提下，将手语者的外观迁移到其他骨架姿态上的方法。通过估计的骨骼姿态，将一名手语者的外观转移到另一名手语者身上，同时保持自然的动作与过渡效果。该方法提升了基于姿态的渲染与手语拼接质量，并实现了身份模糊化。实验表明，该方法虽略微损害手语识别性能，但显著降低了手语者身份识别准确率，揭示了隐私保护与功能实用性之间的权衡。代码已开源：https://github.com/sign-language-processing/pose-anonymization。
原文摘要 · Abstract (English)
We introduce a method for transferring the signer's appearance in sign language skeletal poses while preserving the sign content. Using estimated poses, we transfer the appearance of one signer to another, maintaining natural movements and transitions. This approach improves pose-based rendering and sign stitching while obfuscating identity. Our experiments show that while the method reduces signer identification accuracy, it slightly harms sign recognition performance, highlighting a tradeoff between privacy and utility. Our code is available at https://github.com/sign-language-processing/pose-anonymization.
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。