Deidentification of Drivers’ Face Videos: Scope and Challenges in Human Factors Research

Abstract

Data sharing across disciplines helps to build collaboration, and advance research. With recent development in data-driven models, there is an unprecedented need for data. However, data collected from human research subjects are required to follow proper ethical guidelines. Researchers have an obligation to protect the privacy of research participants and address ethical and safety concerns when data contains personally identifying information (PII). This paper addresses this problem with a focus on sharing drivers’ face videos for transportation research. The paper first gives an overview of the multitude of problems that are associated with sharing drivers’ videos. Then it demonstrates the possible directions for data sharing by de-identifying drivers’ faces using artificial intelligence-based techniques. The results achieved through the proposed techniques were evaluated qualitatively and quantitatively to prove the validity of the suggested methods. We specifically demonstrated how face-swapping algorithms can effectively de-identify faces while still preserving important attributes related to human factor research including eye movements, head movements, mouth movements, etc. Finally, we discuss possible measures to share such de-identified videos with the greater research community.