| Distance teaching satisfies our needs on continuing education and lifelong education, as it is not subject to the limitation of time and geography. However, there are two challenges in the development of distance teaching:the lecture video resources are limited and interaction is poor in existing distance teaching systems. Hence, it is of great significance to develop efficient coding algorithms for interactive applications with high-spatial-resolution videos.Based on the analysis on the advantages and disadvantages of existing distance teaching systems, this thesis proposes the interactively region of interest (ROI) based video coding scheme. Interactively ROI playing is achieved by first extracting the ROIs from high-spatial-resolution videos, and then storing and transferring the video stream of individual ROIs. Evaluation results indicate that our scheme maintains the high quality of the original video, and incurs little overhead on network traffic and computational complexity on clients.On the other hand, encoding complexity is increased as the coding algorithm is applied to multiple ROIs. To make our scheme practical in real applications, we optimize the H.264encoding module that highly degrades system performance. The optimization lies in three aspects:the implementation algorithm of encoder, the parallel optimization of encoder and the optimizations based on TileGx36platform. According to the testing results of parallel design, the running efficiency is improved greatly. |
PDF Full Text Request