臺灣博碩士論文加值系統

基於可調視訊編碼(SVC)之架構，本論文闡釋一個在使用多層編碼控制下，進行編碼速度優化的問題。傳統由下往上的編碼控制在相對於單層編碼上會有不對稱的編碼效率損失。因此為了能夠在基層和增進層之間的編碼效率上做權衡，多層編碼控制技術已在先前被提出來。然而在現今的方法上，有兩個主要的問題存在:(1)基層使用權重式的Lagrangian 成本決策方法。(2)增進層則是使用單層編碼決策方法。前者在解決限制式最佳化問題上，其目標函數和限制條件兩者都將會隨著權重因子的選擇而有所變化。此外又因為後者方法不一致的影響，導致了在增進層上發生了無法預期的結果。為了解決這些問題，我們重提了多層編碼控制的公式化問題。並且因為多層編碼控制存在著編碼速度過於緩慢的缺陷，本文利用主導性配對的觀念以及對增進層決策的重新審視，提出了一個兩階段式的快速決策演算法。實驗結果顯示，本文提出的快速決策演算法跟徹底式搜尋所需的56組配對相比，我們平均僅需要測試13組即可，在這樣的決策組數降低下，編碼平均速度不僅超越徹底式搜尋逾84%，更在編碼品質上沒有太多的失真。此外，產生出來的實驗結果也顯示，新的快速決策演算法相較於過去的決策演算法上，本方法在給予不同的權重因子中更具有可預測性以及連貫性。

This thesis addresses the problem of performing fast mode decision for SVC multi-loop encoder control. The conventional bottom-up encoder control is characterized by its uneven distribution of rate-distortion loss relative to single-layer coding. For a tradeoff between the coding efficiency of the base layer (BL) and the enhancement layer (EL), a multi-layer encoder control was proposed. The current approach, however, poses two major problems: (1) it uses the weighted Lagrangian cost as the search criterion for mode decision at the BL and (2) it adopts the single-layer decision criterion for the EL. The former amounts to solving a constrained optimization problem in which both the objective function and the constraints may vary with the choice of the weighting factor, while the latter can sometimes lead to unpredictable results at the EL. To solve these problems, we have revisited the problem formulation of multi-layer encoder control, and have proposed an improved two-stage algorithm by using the concept of dominant mode pairs and by revising the mode decision criterion at the EL. Experimental results show that our fast mode decision algorithm, on average, needs to check only 13 mode pairs, compared to 56 required for the exhaustive search. The mode set reduction leads to a considerable time saving of 84-88%, with an ignorable change in R-D performance. Besides, the results produced with the new decision criterion are more predictable and consistent
with different choices of the weighting factor.

Contents i
List of Tables iii
List of Figures v
1 Introduction 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Problem Statement . . . . . . . . . . . . . . . . . . . . . . .2
1.3 Contributions . . . . . . . . . . . . . . . . . . . . . . . . .3
1.4 Organization . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Scalable Video Coding and its Encoder Control 5
2.1 Introduction to SVC . . . . . . . . . . . . . . . . . . . . . .5
2.1.1 Concept . . . . . . . . . . . . . . . . . . . . . . . . . . .5
2.1.2 Inter-Layer Prediction . . . . . . . . . . . . . . . . . . . 6
2.2 SVC Encoder Control . . . . . . . . . . . . . . . . . . . . . .8
2.2.1 Bottom-up Encoder Control . . . . . . . . . . . . . . . . . .9
2.2.2 Multi-layer Encoder Control . . . . . . . . . . . . . . . . .10
2.3 Comparison and Summary . . . . . . . . . . . . . . . . . . . . 14
3 Fast Mode Decision for Multi-layer Encoder Control 16
3.1 Proposed Multi-layer Encoder Control . . . . . . . . . . . . . 16
3.2 Analysis of Mode Distribution . . . . . . . . . . . . . . . . .18
3.2.1 Combined Mode Pair Distribution in MLEC . . . . . . . . . . .18
3.2.2 Analyses of Cross-Layer Mode Decision . . . . . . . . . . . . .23
3.3 Fast Mode Decision Algorithm . . . . . . . . . . . . . . . . . 26
3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . .28
4 Experiments and Analyses 30
4.1 Implementation and Test Conditions . . . . . . . . . . . . . . 30
4.2 R-D Tradeoff with Fixed QP Setting . . . . . . . . . . . . . . 31
4.2.1 Schwarz's MLEC . . . . . . . . . . . . . . . . . . . . . . . 31
4.2.2 Proposed MLEC . . . . . . . . . . . . . . . . . . . . . . . . 36
4.3 Performance Evaluation with VariableQP Setting . . . . . . . . 39
4.3.1 Comparisons in Different Criteria . . . . . . . . . . . . . .39
4.3.2 Schwarz's MLEC versus Proposed MLEC . . . . . . . . . . . . . 41
4.4 Performance Comparisons . . . . . . . . . . . . . . . . . . . .42
4.4.1 R-D Performance . . . . . . . . . . . . . . . . . . . . . . .42
4.4.2 Computational Complexity. . . . . . . . . . . . . . . . . . .45
5 Conclusions 48
Bibliography 51

[1] H. Schwarz, D. Marpe, and T. Wiegand, “Overview of the Scalable Video Coding Extension of the H.264/AVC Standard,” IEEE Trans. on Circuits and Systems for Video Technology, vol. 17, September 2007.
[2] J. Reichel, H. Schwarz, and M. Wien, “Joint Scalable Video Model 10 (JSVM 10),” Joint Video Team, JVT07/W202, April 2007.
[3] H. Schwarz and T. Wiegamd, “R-D Optimized Muti-layer Encoder Control for SVC,” Proceeding of Int’l Conference on Image Processing, September 2007.
[4] T. Wiegand, H. Schwarz, A. Joch, F. Kossentini, and G. J. Sullivan, “Rate-Constrained Coder Control and Comparison of Video Coding Standards,” IEEE Trans. on Circuits and Systems for Video Technology, vol. 13, pp. 688—703, July 2003.
[5] D. S. Taubman and M. W. Marcellin, “JPEG2000 Image Compression, Fundamentals, Standards and Practice,” April 2002.
[6] Z. G. Li, S. Rahardja, and H. Sun, “Implicit Bit Allocation for Combined Coarse Granular Scalability and Spatial Scalability,” IEEE Trans. on Circuits and Systems for Video Technology, vol. 16, pp. 1449—1458, December 2006.
[7] T. France, H. Schwarz, and M. Wien, “JSVM Software (Version 9.12.2),” Joint Video Team, JVT08/Z203, January 2008.
[8] X. Li, P. Amon, A. Hutter, and A. Kaup, “Lagrange Multiplier Selection for Rate-Distortion Optimization in SVC,” Joint Video Team, JVT09/AD021, January 2009.
[9] K. Ramchandran, A. Ortega, and M. Vetterli, “Bit Allocation for Dependent Quantization with Applications to Multiresolution and MPEG Video Coders,”IEEE Trans. on Image Processing, vol. 3, pp. 533—545, September 1994.
[10] S. Liu and C.-C. J. Kuo, “Joint Temporal-Spatial Bit Allocation for Video Coding With Dependency,” IEEE Trans. on Circuits and Systems for Video Technology, vol. 15, pp. 15—26, January 2005.