Main content

Quality Metrics in Wireless Video Streaming

Show full item record

Title: Quality Metrics in Wireless Video Streaming
Department: Department of History
Program: Engineering
Advisor: Dony, Robert
Abstract: The advent of efficient video compression standards and the rapid evolution of the wireless networks have made it possible to provide high data rate services to end users. In this context the assessment of the digital video quality has become an increasingly important aspect. Visual assessment techniques are good to predict quality but they are complex, expensive and time consuming. Current mathematical, objective methods measure quality using three different approaches; Full Reference (FR), Reduced Reference (RR) and No Reference (NR). The current existing metrics do not adapt well to wireless environment limitations. HVS based approaches are complex while FR based approaches have limitations due to bandwidth constraints. The NR approach is a distortion specific approach. There are spatial, temporal and network distortions when the video is transmitted over wireless networks, and all of these distortions need to be estimated for an end-to-end quality evaluation, as true quality cannot be predicted by just measuring one kind of distortion. This research proposes a framework to evaluate the video quality which is based on the integration of RR and NR methods by taking into account major spatial, temporal and network impairments along with HVS effects. The framework works with two different kinds of received data. The framework can operate in three modes, i.e., 1) only RR measurement , 2) Hybrid of RR and NR (network, temporal, blurriness, blockiness and HVS) functional blocks, 3) Only NR estimation. The framework has been tested with a experimental set up for video quality evaluation. The JM 16.2 reference codec has been integrated with OMNeT++ simulator for the transmission of actual video packets in real time rather than working off-line. Subjective test data base is also developed using the experimental set up as explained. Framework is verified with this test data base and also verified with H.264 compressed data base provided by the LIVE Video Quality Database. Each individual metric has been tested/verified independently along with the overall testing/verification of the whole framework. Both proposed metrics individually and framework correlate strongly with subjective scores. The framework achieves high performance, scalability and feasibility to assess the video quality.
Date: 2014-09

Files in this item

Files Size Format View
Uzair_Muhammad_201409_Phd.pdf 35.64Mb PDF View/Open

This item appears in the following Collection(s)

Show full item record