| Video streaming technologies are very important in many applications such as web video, Internet news broadcasting, video-on-demand, etc. Since the Internet is a best-effort network, the quality of the streamed video can not be guaranteed. Therefore, how to improve the quality of service (QoS) of video streaming has become an important research topic. Currently, research effort is primarily focused on the streaming of the MPEG/H.26x video which is commonly supported in commercial products. In comparison, three-dimensional (3-D) wavelet compressed video has not been widely used in streaming applications, but its advantages such as great scalability and error resilience ability have attracted the attention of researchers. Although the long initial delay in 3-D compression makes it unsuitable for interactive real-time applications such as videoconference, we see feasibility to use it in stored video delivery which has more relaxed delay constraint. The objective of our research is to improve both the efficiency and the QoS of the 3-D wavelet video streaming. This dissertation consists of two major topics: a highly efficient transport protocol (AVP) for video streaming and a content-based retransmission mechanism. AVP adopts a novel audio-based synchronization scheme to reduce both the header size and the packet rate while achieving implicit synchronization. As a result, the transmission efficiency is improved and the end-to-end message delay is reduced. Furthermore, AVP provides adaptive QoS by a prioritized packetization scheme. The content-based retransmission mechanism is aimed to increase the error recovery ability of the streaming system. The fundamental idea is to consider the content of the packets and the traffic of the Internet simultaneously to optimize the QoS. In the encoding stage, the content information is generated and stored based on which an intelligent hybrid decision system is designed to perform a two-step optimization. First, the bits used for transmitting each group of video frames (GOF) is dynamically allocated according to the changing traffic conditions such that the quality of video is consistent. Secondly, an intelligent scheme for retransmitting the lost packets is designed to minimize the distortion of video. Based on the research, a new framework is designed to achieve effective and efficient video streaming over lossy networks. |
