丁锐, 邝继顺, 谭论正. 用于运动估计的高效三步法的VLSI实现[J]. 武汉大学学报 ( 信息科学版), 2006, 31(9): 842-845.
引用本文: 丁锐, 邝继顺, 谭论正. 用于运动估计的高效三步法的VLSI实现[J]. 武汉大学学报 ( 信息科学版), 2006, 31(9): 842-845.
DING Rui, KUANG Jishun, TAN Lunzheng. VLSI Implementation of Efficient Three-Step Search Algorithm for Motion Estimation[J]. Geomatics and Information Science of Wuhan University, 2006, 31(9): 842-845.
Citation: DING Rui, KUANG Jishun, TAN Lunzheng. VLSI Implementation of Efficient Three-Step Search Algorithm for Motion Estimation[J]. Geomatics and Information Science of Wuhan University, 2006, 31(9): 842-845.

用于运动估计的高效三步法的VLSI实现

VLSI Implementation of Efficient Three-Step Search Algorithm for Motion Estimation

  • 摘要: 根据高效三步法的特点,提出了有效的数据流和相应的VLSI结构。该数据流充分利用参考帧搜索点之间数据重叠的特征,以减少数据存储器访问的次数,从而减少了运动估计中功率消耗最大部分的操作。通过对各种测试序列的仿真证明,该VLSI结构有效地实现了高效三步法,适合MPEG-4和HDTV应用。

     

    Abstract: An efficient data flow pattern and corresponding hardware architecture for efficient three-step search(E3SS) algorithm are preseated for motion estimation.The data flow exploits the overlap of reference data among the search points to reduce data memory accesses that are the most power consuming operations in motion estimation.The corresponding hardware architecture implements the search for two different patterns of E3SS efficiently.Simulation results show that the average number of cycles required to make a block matching for different test sequences is 280,which fulfills the speed requirement of MPEG-4 and HDTV standards.

     

/

返回文章
返回