Çoklu-görüntülü içeriğin internet üzerinden güvenli akıtımı için açık standardlar bazlı sunucu-istemci modeli

The Internet Protocol (IP) is the natural choice for 3D video transmission, if we considering the recent success of many voice over IP (VoIP), video over IP, and IPTV applications. However, the choice of IP for the network layer does not make the transmission application design straightforward, since IP only dictates the use of unreliable packet switched networks. Its flexible architecture allows use of various transmission protocols and policies over IP. Therefore, it is the task of the application to implement an efficient scheme that will most effectively manage the available channel throughput. This can only be achieved by understanding the characteristics of each layer in the protocol stack and implementing a cross-layer design approach.This thesis proposes a novel framework for streaming of 3D video based on multi-view video (MVV) representations. The proposed streaming platform is completely open standards based, flexible, and backwards compatible for supporting monoscopic streaming to legacy clients. The MVV in the server is compressed using a simplified form of MVC with negligible loss of compression efficiency and streamed using RTSP, SDP and RTP to clients. Raptor Codes are used for fighting with the possible packet loses at the network layer. The clients can perform basic error concealment to reduce the effects of remaining packet losses and can decode MVC in real-time. The modular client can display decoded 3D content on a multitude of 3D display systems.The proposed streaming platform has been extensively tested over the Internet to find answers to the following questions related to 3DTV transmission over IP: i) What is the required bandwidth for 3D transmission, and what are the parameters that affact this ratio? ii) Is the cost of enabling slice mode justified for achieving better channel throughput? iii) What is the required rate of channel coding against what percentage of packet losses? iv) How does Raptor Codes perform when we compare its redundancy level against alternative strategies such as Multiple Description Coding? v) What are the required modifications over the standard streaming protocols for 3DTV transmission? vi) How can the design take advantage of the multi-core processors? vii) Are there any other parameters that have significant influence on packet loss ratio beside bit-rate of the stream? If there is, do they affect the packetization strategy?