Droppix: Towards More Realistic Video Fingerprinting

We present preliminary results for a way of video fingerprinting that is different from typical methods based on paradigms from watermarking of still images. Our approach is based on modifying some fragments of the clip in a carefully chosen manner. We show the robustness of our approach against a number of typical of attacks. The marks introduced by our modifications are almost imperceptible to the viewer and their impact can be adjusted. Finally, our protocol is computationally light and can be combined with others schemes as an extra security layer.