Adaptive Carving Method for Live FLV Streaming

Haidong Ge, Ning Zheng, Lin Cai, Ming Xu, Tong Qiao, Tao Yang, Jinkai Sun, Sudeng Hu

Abstract: Currently, most video carving methods are to recover video files from disk file system, but these methods often do not work well for video form network streams, especially for live streaming video. In this paper, an adaptive video carving method is proposed to recover the live FLV (Flash Video) streaming video from network traffic. Firstly, to recover videos when there is no packet loss during data capture, a method based on network data structure is proposed. Secondly, to solve the problem of packet loss or corruption during data capture, another video carving method is proposed based on both the FLV structure and network data structure. Finally, to achieve good balance between computational complexity and recovery accuracy, an adaptive method based above two methods is proposed. The experimental results show that the proposed methods achieve good performance both in consuming time and recovery rate.