Effect of Face Blurring on Human Pose Estimation: Ensuring Subject Privacy for Medical and Occupational Health Applications

• Published in: Sensors (Basel, Switzerland) Vol. 22; no. 23; p. 9376
• Main Authors: Jiang, Jindong, Skalli, Wafa, Siadat, Ali, Gajny, Laurent
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 01.12.2022; MDPI
• Subjects: Algebra; Algorithms; Artificial Intelligence; Asymptomatic; Biomechanical Phenomena; Biomechanics; Cameras; Computer Science; Computer vision; Datasets; Deep learning; Engineering Sciences; Experiments; face blurring; human pose estimation; Humans; kinematic analysis; Kinematics; Machine Learning; Machine vision; Mechanics; Neural networks; Neural Networks, Computer; Occupational Health; Privacy; Privacy, Right of; Statistical significance; deep learning; face blurring; human pose estimation; kinematic analysis; Atomic and Molecular Physics; Deep learning; Analytical Chemistry; Electrical and Electronic Engineering; Human pose estimation; Face blurring; Instrumentation; Biochemistry; Kinematic analysis
• ISSN: 1424-8220; 1424-8220
• DOI: 10.3390/s22239376

The face blurring of images plays a key role in protecting privacy. However, in computer vision, especially for the human pose estimation task, machine-learning models are currently trained, validated, and tested on original datasets without face blurring. Additionally, the accuracy of human pose estimation is of great importance for kinematic analysis. This analysis is relevant in areas such as occupational safety and clinical gait analysis where privacy is crucial. Therefore, in this study, we explore the impact of face blurring on human pose estimation and the subsequent kinematic analysis. Firstly, we blurred the subjects’ heads in the image dataset. Then we trained our neural networks using the face-blurred and the original unblurred dataset. Subsequently, the performances of the different models, in terms of landmark localization and joint angles, were estimated on blurred and unblurred testing data. Finally, we examined the statistical significance of the effect of face blurring on the kinematic analysis along with the strength of the effect. Our results reveal that the strength of the effect of face blurring was low and within acceptable limits (<1°). We have thus shown that for human pose estimation, face blurring guarantees subject privacy while not degrading the prediction performance of a deep learning model.