视频编码并行设计与高效实现关键技术研究

项目名称： 视频编码并行设计与高效实现关键技术研究

项目编号： No.61201238

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 电子学与信息系统

项目作者： 杨志钢

作者单位： 哈尔滨工程大学

项目金额： 24万元

中文摘要： 视频编码技术不断发展进步，编码算法复杂度随之大幅度增加，使得高性能视频编码器的设计实现成为重要的研究问题。本项目围绕视频编码并行设计与高效实现这一主题展开，即要提高编码速度，又要保证压缩效率。为此本项目对视频编码中的两个关键部件- - 环路滤波和运动估计模块进行深入研究，并从多处理单元协同工作的角度对并行编码结构进行探讨。针对环路滤波模块，通过二次流水技术、掩码操作、对处理方式操作和二级片内存储结构设计，解决大量自适应条件跳转不可避免地出现在算法内层循环中的问题，实现软件流水方案。针对运动估计模块，通过加速全局采样搜索并引入临时替代策略，在不减少搜索点数量的情况下实现大范围、变块大小运动估计的高密集度软件流水设计。针对并行编码结构，分析细粒度宏块级并行编码方式的结构特点、加速系数、数据通信频率，并通过建立效力估计模型，研究编码流程、并行结构、图像内容差异以及平台因素对宏块级并行编码效力的影响。

中文关键词： 视频编码；环路滤波；运动估计；模式决策；并行编码

英文摘要： Due to the continuous development of video coding techniques, the computational complexity of video coding algorithms is rapidly increasing, especially in high resolution applications. Thus, how to efficiently implement a high performance video encoding system becomes an important issue. In this paper, we focus on the techniques for parallel design and efficient implementation of video encoding, and try to achieve both high encoding speed and high coding efficiency. To reach this goal, two key units in the video encoder which are in-loop deblocking filter and motion estimation, and parallel video encoding structures are carefully studied. Firstly, software pipeline solutions to in-loop deblocking filter are presented. Several techniques such as two-pass pipelines, mask operation, pair processing and two-level internal memory organization are adopted to solve the problem that highly adaptive conditional branches inevitably appear in the inner most loops of the algorithm. Then, a high-density software pipeline scheme for large search range and variable block size motion estimation is proposed, which can bring a great running time reduction without decreasing the search points. In the scheme, large search-range motion estimation is optimized by accelerating the global search, and the temporary replacement strategy

英文关键词： video coding；in-loop deblocking filter；motion estimation；mode decision；parallel encoding