系统工程理论与实践
繫統工程理論與實踐
계통공정이론여실천
Systems Engineering—Theory & Practice
2014年
12期
3221~3237
,共null页
精细可扩展编码 贪婪算法 视频流媒体系统 基本层速率
精細可擴展編碼 貪婪算法 視頻流媒體繫統 基本層速率
정세가확전편마 탐람산법 시빈류매체계통 기본층속솔
fine grained scalable coding; greedy algorithm; streaming media system; base layer rate
精细可扩展（fine grained scalable,FGS）编码是网络视频中最流行的编码技术之一.采用FGS编码时,基本层的速率选择需要在受益客户数量和视频编码效率之间进行折中.为此,本文提出了一个源端基本层最佳速率确定算法,以最大化所有客户的观感体验.论文首先提出了一个通过实验和曲线拟合建立R-Q（Rate-Quality）函数曲线的方法,在此基础上给出了缩小最佳基本层速率搜索空间的定理,并针对单序列和多序列的不同特点和不同的问题复杂度,提出了不同的算法.针对单视频序列优化问题,提出了改进的FGSOPT_N法,算法只用到了序列的R-Q函数曲线;针对多视频序列问题是NP困难问题的实际,提出了改进的分枝定界MFGSOPT_N算法和改进的贪婪算法MFGS_N以确保找到最优基本层速率,减少算法运行时间.针对每种算法,给出了实例进行算法效果的演示和对比.结果显示提出的算法能可靠找到最优基本层速率.
精細可擴展（fine grained scalable,FGS）編碼是網絡視頻中最流行的編碼技術之一.採用FGS編碼時,基本層的速率選擇需要在受益客戶數量和視頻編碼效率之間進行摺中.為此,本文提齣瞭一箇源耑基本層最佳速率確定算法,以最大化所有客戶的觀感體驗.論文首先提齣瞭一箇通過實驗和麯線擬閤建立R-Q（Rate-Quality）函數麯線的方法,在此基礎上給齣瞭縮小最佳基本層速率搜索空間的定理,併針對單序列和多序列的不同特點和不同的問題複雜度,提齣瞭不同的算法.針對單視頻序列優化問題,提齣瞭改進的FGSOPT_N法,算法隻用到瞭序列的R-Q函數麯線;針對多視頻序列問題是NP睏難問題的實際,提齣瞭改進的分枝定界MFGSOPT_N算法和改進的貪婪算法MFGS_N以確保找到最優基本層速率,減少算法運行時間.針對每種算法,給齣瞭實例進行算法效果的縯示和對比.結果顯示提齣的算法能可靠找到最優基本層速率.
정세가확전（fine grained scalable,FGS）편마시망락시빈중최류행적편마기술지일.채용FGS편마시,기본층적속솔선택수요재수익객호수량화시빈편마효솔지간진행절중.위차,본문제출료일개원단기본층최가속솔학정산법,이최대화소유객호적관감체험.논문수선제출료일개통과실험화곡선의합건립R-Q（Rate-Quality）함수곡선적방법,재차기출상급출료축소최가기본층속솔수색공간적정리,병침대단서렬화다서렬적불동특점화불동적문제복잡도,제출료불동적산법.침대단시빈서렬우화문제,제출료개진적FGSOPT_N법,산법지용도료서렬적R-Q함수곡선;침대다시빈서렬문제시NP곤난문제적실제,제출료개진적분지정계MFGSOPT_N산법화개진적탐람산법MFGS_N이학보조도최우기본층속솔,감소산법운행시간.침대매충산법,급출료실례진행산법효과적연시화대비.결과현시제출적산법능가고조도최우기본층속솔.
The fine grained scalable （FGS） coding is one of the most popular coding technologies in network video stream. In FGS stream, the base layer rate is an important parameter and should generally make a compromise between the number of beneficial clients and the efficiency of video coding. To maximize all clients＇ perceived quality, the selection algorithm for the optimal base layer rate is proposed. This paper first puts forward a method to build the R-Q （Rate-Quality） function curve and proposes an algorithm to cut search space of the base layer rate. Second, because the different feature and different problem complexity of single and multiple sequences, the different algorithms are proposed. According to the optimization problem of single sequence, the improved FGSOPT_N algorithm is proposed, which only bases on R-Q function curve of a sequence and according to the fact that the optimization problem of multiple sequences is NP hard~ the improved MFGSOPT_N algorithm and greedy algorithm MFGS_N are proposed to improve the probability to find the optimal base layer rate, thus to reduce the runtime. At last, for each algorithm, illustration examples are given to demonstrate and compare the effect of algorithm. The results show the algorithms can reliably find the optimal base layer rate if it exits.
