Wireless device-to-device (D2D) communication has empowered efficient and convenient video sharing among neighboring devices. However, the mobility and diversity of user devices, combined with dynamically shifting channel conditions, make these processes susceptible to environmental interference. In this paper, we introduce an adaptive network coding scheme tailored for D2D raw source video streaming in the scalable video coding (SVC) format. Video frames are partitioned into equal-sized fragments and encoded into streaming packets with the addition of certain redundancy. To address the observed deficiency where decoding throughput of coded video frames is generally lower than that of the encoding operation on mobile device platforms, we introduce a parallel decoding algorithm based on LU decomposition. Furthermore, we propose a flexible adjustment mechanism, named Nefis, which dynamically adapts the network coding redundancy and stream resolution based on the statistical model established to predict current network conditions and video streaming quality. Implementation on Android platform demonstrates that Nefis reduce redundancy in bandwidth usage during streaming process and enhance resilience to network dynamics. Experimental results also conclusively demonstrate the feasibility and advantages of network coding technology in D2D streaming applications. Nevertheless, achieving these advantages requires carefully designed encoding and decoding mechanisms.

中文翻译：

---

Nefis：一种基于网络编码的灵活的设备到设备视频流方案


无线设备到设备（D2D）通信使相邻设备之间能够高效便捷地共享视频。然而，用户设备的移动性和多样性，加上动态变化的信道条件，使得这些过程容易受到环境干扰。在本文中，我们介绍了一种针对可扩展视频编码（SVC）格式的 D2D 原始源视频流定制的自适应网络编码方案。视频帧被分割成大小相等的片段，并编码成流数据包，并添加一定的冗余。为了解决编码视频帧的解码吞吐量通常低于移动设备平台上的编码操作的缺陷，我们引入了一种基于 LU 分解的并行解码算法。此外，我们提出了一种灵活的调整机制，名为Nefis，它根据为预测当前网络状况和视频流质量而建立的统计模型动态调整网络编码冗余和流分辨率。 Android 平台上的实施表明 Nefis 减少了流媒体处理过程中带宽使用的冗余，并增强了对网络动态的弹性。实验结果也最终证明了网络编码技术在D2D流媒体应用中的可行性和优势。然而，实现这些优势需要精心设计的编码和解码机制。