Adapting the streaming video based on the estimated motion position

Abstract

In real time video streaming, the frames must meet their timing constraints, typically specified as their deadlines. Wireless networks may suffer from bandwidth limitations. To reduce the data transmission over the wireless networks, we propose an adaption technique in the server side by extracting a part of the video frames that considered as a Region Of Interest (ROI), and drop the part outside the ROI from the frames that are between reference frames. The estimated position of the selection of the ROI is computed by using the Sum of Squared Differences (SSD) between consecutive frames. The reconstruction mechanism to the region outside the ROI is implemented in the mobile side by using linear interpolation between reference frames. We evaluate the proposed approach by using Mean Opinion Score (MOS) measurements. MOS are used to evaluate two scenarios with equivalent encoding size, where the users observe the first scenario with low bit rate for the original videos, while for the second scenario the users observe our proposed approach with high bit rate. The results show that our technique significantly reduces the amounts of data are streamed over wireless networks, while the reconstruction mechanism will provides acceptable video quality.