Research On Uncoded Wireless Transmission For 3D Video

With the development of communication and Internet technologies,people's demand for high-quality video is increasing.The two-dimensional(2D)video can no longer satisfy the people's visual experience,and the three-dimensional(3D)video technology which can provide stereoscopic sense has became a research hotspot in the current video field.Threedimensional video applications based on mobile terminals also show great potential for development.Due to the dramatic changes in the state of the wireless channel,the transmission schemes based on source-channel-coding in existing technologies have been unable to satisfy broadcast and mobile scenarios.Recently,a novel uncoded scheme called SoftCast has been proposed for wireless video transmission,which eliminates the cliff effect by traditional coding frameworks and achieves linear quality transition within a wide range of channel Signal-to-Noise Ratio(SNR).Unfortunately,it cannot be directly applied to 3D video wireless transmission,and existing SoftCast-based 3D video research is also rare.Therefore,this thesis is based on how to effectively apply the SoftCast framework to 3D video systems.The specific research work is as follows:1.A key issue in SoftCast-based 3D video transimission is how to allocate a limited power budget to the texture videos and depth maps to get the best overall quality at the receiving end.Therefore,A power-distortion based joint power allocation scheme between texture and depth is proposed to deal with the optimal power allocation problem.Simulation results show that the method is close to the full search algorithm.Compared with the full search algorithm,the complexity of the proposed algorithm is negligible.In addition,as compared with the fixed 1:1 ratio power allocation method between texture and depth,the proposed method can achieve a maximum 1.8dB gain.2.The depth map of the 3D video shows a completely different feature than the texture video.Aiming at the problem of depth map transmission,an uncoded wireless transmission strategy which based on block DCT(Discrete Cosine Transform)for depth map is proposed to deal with it.The strategy has three custom optimizations that take full advantage of the depth map segmentation smoothing feature,while also taking advantage of the fact that the depth map is only used to synthesize virtual viewpoints rather than watch directly.The experimental results show that compared with the traditional transmission scheme,the proposed scheme can provide better performance than the existing SoftCast.3.Multi-view 3D video has a lot of redundancy between different viewpoints,and there is also a certain correlation between depth map and texture video.Aiming at this phenomenon,a joint texture-and-depth transmission strategy for 3D video,which is suitable for multi-view video wireless transmission,is proposed to improve overall video quality.The strategy has four custom optimizations that significantly reduce the average energy of the block.Simulation results show that this scheme can significantly improve the performance of multi-view video SoftCast.