视频传输跨层前向纠错技术分析.pdfVIP

ABSTRACT Withtherapiddevelopmento mobileInternetandtheintelligentterminals,there is an increasing demand for the real-time transmission o video over wireless networks. In wireless networks, video transmission quality is usually a ected by obstacles, surroundings, weather and signal attenuation, leading to severe packet losses. Forward error correction addsredundancy to the source information, and the receiver canusethisredundancytorecoverthe lostdata. Sinceawirelessnetwork is o limited bandwidth, the reasonable allocation o bandwidth is needed for further improvemento videoqualityatthereceiver. Today’smost widely used video coding standardH.264 andthenew generation coding standard H.265 aim to maximize compression e iciency through predictive coding, which creates strong spatio-temporal dependency and di erent priorities among di erent video packets. The loss o di erentvideo packets leads to di erent distortion in the reconstructed video, thus FEC-based video transmission is used in conjunctionwithunequalerrorprotectiontoachievebettervideoquality. Cross-layer Forward error correction can allocate the limited bandwidth more e iciently, which has been widely applied to video transmission over error-prone wireless channels. In this paper, a near-optimal cross-layer forward error correction scheme is presented for H.264/AVC video transmission with channel bandwidth constraints over wireless networks, where rate compatible punctured convolutional codesareusedatthephysicallayerandRaptorcodesareusedattheapplicationlayer. Firstly, inorderto fullyexploittheunequal importanceo thecompresseddata, each source packet is assigned a di erent priority according to its contribution to the reconstructed video quality. Then, the op