Blur Filtration Fails to Preserve Privacy for Home-Based Video Conferencing (22 Oct 2004)
Always-on video provides rich awareness for co-workers separated by distance, yet it has the potential to threaten privacy as sensitive details may be inadvertently broadcasted to others. This threat increases for telecommuters who work at home and connect to office-based colleagues using video because the once private home is now made public. Researchers have looked at video obfuscation techniques as a means to balance privacy and awareness in video conferencing systems. One such technique is blur filtration, which blurs video to hide sensitive details while still giving the viewer a sense of what is going on. While other researchers found that blur filtration mitigates privacy concerns in low-risk office settings, we do not know if it works for riskier situations that can occur in telecommuting settings. Using a controlled experiment, we evaluated blur filtration for its effectiveness in balancing privacy with awareness for typical home situations faced by telecommuters. Participants viewed five video scenes containing a telecommuter at ten levels of blur, where scenes ranged from little to extreme privacy risk. They then answered awareness and privacy questions about these scenes. Our results show that blur filtration is unable to balance privacy with awareness for home scenarios exhibiting even minimally risky situations. Subjective reactions by study participants to video obfuscation as a means to mitigate risk suggest that other image processing techniques will be problematic as well. The implication to media space design is that, by themselves, image processing techniques will not suffice for privacy protection in video-based telecommuting situations. We argue that other context-aware privacy-protecting strategies are required to complement image manipulation and present a prototype design of a context-aware home media space.
Article URL: http://grouplab.cpsc.ucalgary.ca/papers/2005/05-BlurFiltration.TOCHI/abstract.html
Read 37 more articles from GroupLab sorted by
Next Article: Enabling Informal Communication of Digital Stories