Computer Vision News - June 2016

Adam Harvey is a Berlin-based artist, designer and researcher. His work on computer vision and image processing technologies is completely different from what we are used to witness in our industry: Harvey explores the impacts of power asymmetries in new technologies, mostly related to privacy and self-protection from a culture of surveillance . He has initiated camouflage and face-detection spoofing projects like CV Dazzle and Camoflash in order to thwart those cameras which, in his own words, have metamorphosed from art making tools into enablers of surveillance societies. Some of our readers might disagree with Harvey’s stand, but discussing with him is a truly fascinating experience and we are glad to share it with our readers. Computer Vision News: What is it that some computers and cameras are doing and that we should hide from? Why should we be worried about it? Adam Harvey: I think it is important to phrase that question in the right terms. When you say that people are hiding from something, that assumes that the observation activity is legitimate. It thus suggests that hiding is a defensive strategy. I think that point of view can be problematic, since it encourages people to think that hiding is a subversive behavior. On the contrary, I think it should be seen as a way for people to control their visibility . The first step of controlling your visibility is being aware and understanding the ways you are being observed, what is used of the information which is being collected, how is this information processed and how does that affect you, at a benefit or at a cost for your privacy. Are you able to control or modulate your visibility to serve your interests or possibly those of your business? When I allow any computer algorithm or person to observe me and to process me in an algorithmic way, I need to know what that algorithm is doing, be comfortable with it and be sure that it is used for something that is ethical (knowing that sometimes ethics butt heads with a bottom line). BASIS. Frankfurt 2015 8 Computer Vision News Guest Adam Harvey Camoflash detects a camera flash and responds in less than .1 milliseconds with a 12,000 lumen LED pulse that overexposes the targeted camera’s sensor [Anti-Paparazzi Clutch Prototype #2. Photo: Adam Harvey]