Research On Cross-layer Forward Error Correction For Video Transmission

With the rapid development of mobile Internet and the intelligent terminals,there is an increasing demand for the real-time transmission of video over wireless networks.In wireless networks,video transmission quality is usually affected by obstacles,surroundings,weather and signal attenuation,leading to severe packet losses.Forward error correction adds redundancy to the source information,and the receiver can use this redundancy to recover the lost data.Since a wireless network is of limited bandwidth,the reasonable allocation of bandwidth is needed for further improvement of video quality at the receiver.Today's most widely used video coding standard H.264 and the new generation coding standard H.265 aim to maximize compression efficiency through predictive coding,which creates strong spatio-temporal dependency and different priorities among different video packets.The loss of different video packets leads to different distortion in the reconstructed video,thus FEC-based video transmission is used in conjunction with unequal error protection to achieve better video quality.Cross-layer Forward error correction can allocate the limited bandwidth more efficiently,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 codes are used at the physical layer and Raptor codes are used at the application layer.Firstly,in order to fully exploit the unequal importance of the compressed data,each source packet is assigned a different priority according to its contribution to the reconstructed video quality.Then,the optimal transmission in the ideal situation,which guarantees the most important packets to be correctly transmitted as many as possible in the ideal situation,is obtained through theoretical analysis.Finally,a heuristic algorithm is developed to get the near-optimal transmission parameters used in the practical situation searching from the optimal solution obtained in the ideal situation.Simulation results show that,compared with the previous work the proposed method can provide a better received video quality.