物联网的基础理论与方法研究.pptVIP

Then we further propose an approach to infer a user’s basic activities of daily living based on only the traffic size data of an encrypted video stream by exploiting the leaked side-channel information. In this work, we only consider the case where only one person appears in the visual view of the camera. Our method consists of three steps. The first step is very straitforward, I will skip this step, you can find the details in our paper. * After that, we split the traffic size data into small pieces with each corresponding to one activity. To split each activity, we first locate the beginning and ending points of each activity. We observe that there will be a change point in the traffic size data when an activity ends and another activity starts. This is because that the statistical properties will change. As shown in the first figure, when dressing ends, and moving starts, the traffic size rises and then drop to a lower level in a short interval. This is because that when dressing ends and moving starts, the motions change substantially. * * * * * * * * * * * * Reducing Temporal Redundancy by Difference Coding * Original frames Difference coding GOPs (Group of Pictures) Traffic size of the encoded video streams Side-Channel Information Leakage When the monitoring area is static, the size of B/P-frame would be very small, and the traffic will be steady and small; When someone appears in the monitoring area, the size of B/P-frame will be relatively large, and the traffic will be large and unstable; The traffic patterns are different when a user performs different activities. Large-scale and complex movements generally yield heavy and unstable traffic caused by the large and varied differences between adjacent frames. Are these claims right? An Experimental Study Experiment settings: Used Google’s Dropcam Pro, and recorded the traffic size data of an encrypted video stream without I frames A volunteer conducted four basic activities of daily living (ADL