About Face: A Survey of Facial Recognition Evaluation
It addresses the need for transparency in facial recognition datasets to improve real-world understanding and mitigate risks, though it is incremental as a survey.
The paper surveys over 100 face datasets from 1976 to 2019, totaling 145 million images of 17 million subjects, to analyze how political, technological, and normative influences shape dataset construction and mask potentially harmful practices.
We survey over 100 face datasets constructed between 1976 to 2019 of 145 million images of over 17 million subjects from a range of sources, demographics and conditions. Our historical survey reveals that these datasets are contextually informed, shaped by changes in political motivations, technological capability and current norms. We discuss how such influences mask specific practices (some of which may actually be harmful or otherwise problematic) and make a case for the explicit communication of such details in order to establish a more grounded understanding of the technology's function in the real world.