Recoverable Anonymization for Pose Estimation: A Privacy-Enhancing Approach
This work addresses privacy concerns in surveillance applications for users and stakeholders by offering a recoverable anonymization method, though it is incremental as it builds on existing privacy-enhancing techniques.
The paper tackles the privacy-performance trade-off in human pose estimation for surveillance by proposing a system that generates privacy-enhanced portraits while maintaining high pose estimation performance, with experimental results showing robust performance in privacy enhancement, sensitive personal information recovery, and pose estimation.
Human pose estimation (HPE) is crucial for various applications. However, deploying HPE algorithms in surveillance contexts raises significant privacy concerns due to the potential leakage of sensitive personal information (SPI) such as facial features, and ethnicity. Existing privacy-enhancing methods often compromise either privacy or performance, or they require costly additional modalities. We propose a novel privacy-enhancing system that generates privacy-enhanced portraits while maintaining high HPE performance. Our key innovations include the reversible recovery of SPI for authorized personnel and the preservation of contextual information. By jointly optimizing a privacy-enhancing module, a privacy recovery module, and a pose estimator, our system ensures robust privacy protection, efficient SPI recovery, and high-performance HPE. Experimental results demonstrate the system's robust performance in privacy enhancement, SPI recovery, and HPE.