Meirose, FranziskaSchultze, SvenKuehlewind, SebastianKoelle, MarionAbdenebaoui, LarbiBoll, SusanneDachselt, RaimundWeber, Gerhard2018-07-092018-07-092018https://dl.gi.de/handle/20.500.12116/16658Talking to each other is personal, maybe even intimate. Thus, privacy expectations are particularly high during interpersonal conversations, and image or audio recordings are problematic in these contexts. In consequence, smart glasses and other body-worn devices with “always-on” cameras are not well accepted during interpersonal conversations. Proposing a simple-to-implement computer vision procedure, we work towards a solution to this issue. Using imagery from a head-worn camera we detect face-to-face conversations in real-time, as well as distinguish between intimate, personal and social conversations based on intrinsic camera parameters. Starting from a fictive scenario, we illustrate how this knowledge can be used for interaction designs that increase both, the users’ as well as their bystanders’ privacy, e.g., by muting audio or disabling the camera. Finally, we suggest directions for future work.enWearable ComputingUbiquitous ComputingPrivacyFace DetectionText/Workshop Paper10.18420/muc2018-mci-0336