A Novel Power Reduction Technique for Block Matching Motion Estimation Hardware
Abdulkadir Akın, Onur Can Ulusel, Tevfik Zafer Özcan, Gokhan Sayilar, İlker Hamzaoğlu · 2011
Motion Estimation (ME) is the most computationally intensive part of video compression and video enhancement systems. Therefore, in this paper, we propose comparison prediction (CP) technique for reducing the power consumption of block matching (BM) ME hardware. CP technique reduces the power consumption of absolute difference operations performed by BM ME hardware. CP technique can easily be used in all BM ME hardware. In this paper, we applied it to a 256 processing element fixed block size ME hardware implementing full search algorithm. It reduced the average dynamic power consumption of this ME hardware by 2.2% with no Peak Signal-to-Noise Ratio (PSNR) loss and by 9.3% with 0.04% PSNR loss on a XC2VP30-7 FPGA.