Smart Privacy Visor: Bridging the Privacy Gap

Abstract

Due to the propagation of devices with imaging capabilities, the amount of pictures taken in public spaces has risen. Due to this, nintentionally photographed bystanders are often represented in pictures without being aware of it. Social networks and search engines make these images easier accessible due to the available meta-data and the tagging and linking functionality provided by these services. Facial recognition amplifies the privacy implications for the individuals in these pictures. Overall there exist three main classes of wearable picture-related Privacy Enhancing Technologies (PETs). As they need different prerequisites to operate and become effective they have unique time frames in the future where they can be effective even if introduced today. The group of face pattern destroying picture PETs work directly against current face detection algorithms and is the choice for immediate usage. These PETs destroy face patterns and inhibit the detection and automated processing and meta-data enrichment of individuals. This unconditionally visual destructive behavior can be a major obstacle in transition to other PETs. In this paper, we describe how to master a smooth transition between these classes including the restoration of the visual damage some of these methods entail. Furthermore, we propose the Smart Privacy Visor, a PET which combines the previously published Privacy Visor and the Picture Privacy Policy Framework. The overall goal of this transition is to create a PET that avoids identifiable and linkable properties which contradicts the goals of picture PETs in the first place and offer a visually appealing photographic result at the same time.