在先前的研究[8]中，畫面基準下的降取樣壓縮編碼提升了H.264/AVC畫面內預測模式效能。我們藉由加入了四分之一降取樣模式和選擇是否加入預測錯誤補償畫面提出了一個畫面基準下最佳適應性降取樣編碼。但是影像畫面會因為畫面內部各區域所含有的特性不同，限制了畫面基準下的降取樣壓縮編碼的效能。因此，我們將降取樣壓縮編碼從畫面基準下應用到區塊基準下，並提出區塊基準下最佳適應性降取樣壓縮編碼。最後實驗結果顯示：相較於先前的研究[8]，我們所提出的演算法在畫面基準下平均增加了0.295%到3.09%的位元率增益；而在區塊基準下平均效能則增加了5.352%的位元率增益。In previous work [8], frame-based down-sampling video coding technique improves efficiency in intra frame of H.264/AVC. An algorithm of frame-based down-sampling video coding is proposed to enhance the efficiency: the quarter mode and not to add prediction error are considered in it. However, frames would limit the efficiency because every region in frame might have a different character. We apply down-sampling video coding technique from frame-based into block-based, proposing a block-based algorithm of down-sampling video coding. The experiments show the frame-based algorithm, comparing with [8], achieves 3.09% of BD-bitrate saving in average, and the block-based algorithm further improves coding performance (5.352% of BD-bitrate saving in average).
