一种基于块匹配算法的SAD运算加速器

J4 ›› 2012, Vol. 34 ›› Issue (7): 60-64.

一种基于块匹配算法的SAD运算加速器

谷会涛，陈书明

(国防科学技术大学计算机学院，湖南长沙 410073）

收稿日期:2010-06-28 修回日期:2010-10-26 出版日期:2012-07-25 发布日期:2012-07-25
基金资助:
国家863计划资助项目（2009AA011704）；教育部“高性能微处理器技术”创新团队研究计划

A BlockMatching Algorithm Based Accelerator for SAD Computation

GU Huitao,CHEN Shuming

(School of Computer Science, National University of Defense Technology,Changsha 410073,China)

Received:2010-06-28 Revised:2010-10-26 Online:2012-07-25 Published:2012-07-25

摘要/Abstract

摘要：

基于块匹配算法的运动估计是图像和视频应用中的关键技术。SAD运算是运动估计中最主要的运算形式，具有极高的计算复杂度和传输带宽需求。本文提出了一种可配置的SAD运算加速器结构，采用一个16×1规模的PE阵列和一个加法树结构加速SAD运算的执行。本文将PE阵列和加法树结构的流水线进行细致划分，有效提高了工作频率。加速器采用DMA事件机制，大部分的数据传输可以与SAD计算并行进行，减少了数据传输延迟引起的性能下降。实验结果显示，搜索16×16大小的搜索窗口，本文结构只需要4 102个周期。基于SMIC 0.13μm的CMOS标准单元工艺对本文结构进行综合，最高工作频率可达到750MHz，面积约为16.8k门和3.5 KB的片上存储器。

关键词: 运动估计, 块匹配算法, 处理单元阵列, 视频编码

Abstract:

Blockmatching based motion estimation is one of the most important techniques in image and video applications. The sum of absolute difference (SAD) is the major computation in motion estimation and requires huge computation complexity and transmission bandwidth. This paper proposes a reconfigurable SAD accelerator, in which a 16×1 processing elements (PE) array and an adder tree structure are used to improve the execution speed of SAD computation. The pipeline partition of PE array and adder tree is performed carefully in order to increase the work frequency. In order to reduce the performance loss caused by data transfer delay, a DMA event mechanism is employed to transmit data when the SAD accelerator is working. The experimental results show that, the proposed architecture needs 4102 cycles for searching a 16×16 search window. With a 0.13μm CMOS standard cell technology, the proposed accelerator requires only 16.8 k gates and 3.5 KB of memory at the 750MHz operation frequency.

Key words: motion estimation;BMA;processing element array;video coding

谷会涛，陈书明. 一种基于块匹配算法的SAD运算加速器[J]. J4, 2012, 34(7): 60-64.

GU Huitao,CHEN Shuming. A BlockMatching Algorithm Based Accelerator for SAD Computation[J]. J4, 2012, 34(7): 60-64.

参考文献

［1］Wiegand T, Sullivan G J, Bjntegaard G, et al. Overview of the H.264/AVC Video Coding Standard［J］. IEEE Transactions on Circuits and Systems for Video Technology, 2003, 13(17):560576.
［2］Jachalsky J, Wahle M, Pirsch P, et al. A Core for Ambient and Mobile Intelligent Imaging Applications［C］∥Proc of IEEE International Conference on Multimedia and Expo, 2003:14.
［3］De Haan G. Video Format Conversion［J］. Journal of Society Information Display, 2000, 8(1):7987.
［4］Ernst F,Wilinski P,Van Overveld K.Dense Structure frommotion: An Approach Based on Segment Matching［C］∥Proc of the European Conference on Computer Vision, Copenhagen, 2002:1723.
［5］Cao Wei, Mao Zhigang, Lu Zhiqiang, et al. VLSI Architecture Design for VariableSize Block Motion Estimation in MPEG4 AVC/H.264［C］∥Proc of the IEEE AsiaPacific Conf on Circuits and Systems, 2004:617620.
［6］Yap S Y, McCanny J V. A VLSI Architecture for Variable Block Size Video Motion Estimation［J］. IEEE Transactions on Circuits and System  II: Express Briefs, 2004, 51( 7): 384389.
［7］Kim M, Hwang I, Chae S I. A Fast VLSI Architecture for FullSearch Variable Block Size Motion Estimation in MPEG4 AVC/H.264 ［C］∥Proc of the Asia and South Pacific Design Automation Conference, 2005:631634.
［8］Cao Wei, Mao Zhigang. A Novel VLSI Architecture for VBSME in MPEG4 AVC/H.264［C］∥Proc of the IEEE International Symposium on Circuits and Systems. Piscataway, 2005:17941797.
［9］邓磊,高文,胡铭曾,等. 基于AVC/AVS标准高效运动估计硬件结构设计［J］. 计算机研究与发展, 2006, 43(11):19721976.
［10］Rabaey J M, Chandrakasan A, Nikolic B. Digital Integrated Circuits: A Design Perspective［M］. Second Edition. New Jersey: Prentice Hall, 2003.

[1]	崔馨月, 蒋林, 杨坤, 惠超, 胡传瞻, 赵静. 一种HEVC帧内预测算法的动态自重构实现方法[J]. 计算机工程与科学, 2022, 44(12): 2120-2127.
[2]	杨坤, 蒋林, 谢晓燕, 邓军勇, 刘新闯, 胡传瞻. HEVC中率失真优化算法的动态可重构实现[J]. 计算机工程与科学, 2021, 43(02): 354-361.
[3]	覃远年，梁仲华. 基于混合粒子群算法的运动估计研究[J]. 计算机工程与科学, 2019, 41(04): 758-764.
[4]	彭磊，彭勇，吴锡生，徐赟，胡旻涛. 基于UMHexagonS的运动估计算法优化[J]. 计算机工程与科学, 2016, 38(10): 2109-2114.
[5]	周全1,2,3,王中元1,2,3. H.264码率控制跳帧算法的优化[J]. J4, 2016, 38(02): 325-330.
[6]	秦龙龙1,钱渊1,侯雪2，张晓燕1. 基于维纳线性预测的Horn-Schunck光流运动矢量优化算法[J]. J4, 2015, 37(07): 1360-1365.
[7]	袁梅冷1，2，杨张2，雷海军2. 一种基于JMVC参考模型的运动估计优化算法[J]. J4, 2015, 37(06): 1208-1213.
[8]	杨琪，李小曼，罗超宇. H.264帧间模式选择快速算法[J]. J4, 2015, 37(05): 993-1000.
[9]	刘治，王玲，陶小娟. 动态模式的快速运动估计算法[J]. J4, 2014, 36(09): 1760-1764.
[10]	叶仕通. 自适应选择编码模式的快速运动估计算法[J]. J4, 2014, 36(09): 1780-1787.
[11]	端木春江，董朵. 新一代高效视频编码标准中帧内预测快速算法研究[J]. J4, 2014, 36(06): 1042-1049.
[12]	王佳利，姜珊，双凯. 基于时空预测向量相关性的运动估计算法[J]. J4, 2014, 36(03): 502-507.
[13]	苏华友，伍楠，文梅，任巨，吴伟，张春元. 高清H.264变换编码的流式实现[J]. J4, 2011, 33(8): 148-153.
[14]	陈天壮1,梁久祯1,韩军2. 一种基于块运动类型和方向预测相结合的快速搜索算法[J]. J4, 2011, 33(7): 112-117.
[15]	崔大为，窦文华，计永兴. 三维视频压缩研究进展[J]. J4, 2011, 33(5): 63-68.