臺灣博碩士論文加值系統

本文提出了可適應搜索範圍（SR）演算法。一個好的搜索範圍能實現出色的編碼性能並降低內存存取頻寬，因此適當的SR大小對視頻編碼是很重要的。為了達到這個目標，我們首先從時間和空間上相鄰塊的運動向量（MVs）獲得運動向量預測（MVP）。根據MVP的準確性，我們能更快速的發現最佳匹配的圖像塊。我們通過利用運動差異（MD），MVP和MV之間的關係，制定出一套可適應搜索範圍演算法。此外，我們結合基於SAD提前終止演算法進一步使結果更完美。實驗結果顯示該演算法的性能不僅與全搜尋演算法非常相似更能顯著地降低計算複雜度。

This paper presents an adaptive search range (SR) selection algorithm in fast motion estimation. Since a good decision may achieve excellent coding performance with reduced memory access bandwidth, a proper SR size is important to video encoding. To attain this goal, we first obtain a motion vector predictor (MVP) is got from the motion vectors (MVs) of spatially and temporally neighboring blocks. Due to the accuracy of the MVP, we will find the best matching image block faster in the full search. By exploiting the relationship between the motion difference (MD), the MVP and the MV, we develop an adaptive SR selection algorithm. In addition, we further improve by combining SAD-based early termination algorithm. Experimental results show that the performance of the proposed algorithm is very similar to that of the full search algorithm it enables significant reduction in the computational complexity.

摘 要 I
ABSTRACT II
誌謝 III
CONTENTS IV
LIST OF TABLES VI
LIST OF FIGURES VII
INTRODUCTION 1
1.1 Research Background 1
1.2 Motivation 4
1.3 Organization for Thesis 5
INTER CODING AND RQT STRUCTURE IN HEVC 6
2.1 Overview of HEVC 6
2.2 Quad Tree Structure in HEVC 9
2.2.1 Coding Unit 9
2.2.2 Prediction Unit 9
2.2.3 Transform Unit 11
2.3 Inter Coding in HEVC 13
2.3.1 Merge Scheme 14
2.3.2 Advanced Motion Vector Prediction (AMVP) 15
THE PROPOSED ALGORITHM 22
3.1 Overview of Proposed Algorithm 23
3.2 Enhanced AMVP Mechanism 25
3.2.1 Analysis of Motion Vector Predictor Candidates 25
3.2.2 Selection of Motion Vector Predictor Candidates 27
3.3 Adaptive Motion Search Range Decision Algorithm 32
3.3.1 Association between MVD and SR 33
3.3.2 Adaptive Motion Search Range Selection 35
3.3.3 SAD-based Early Termination Algorithm 38
SIMULATION RESULTS 42
4.1 Simulation Settings 42
4.2 Simulation Results for Enhanced AMVP Mechanism 48
4.3 Simulation Results for The Proposed Algorithm 50
CONCLUSIONS AND FUTURE WORKS 59
5.1 Conclusions 59
5.2 Future Works 60
REFERENCE 61

[1]K. McCann, W. J. Han, and I. K. Kim, “Samsung’s response to the call for proposals on video compression technology, Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11, Doc. JCTVC-A124, Apr. 2010.
[2]G. J. and J. R. Ohm, “Recent developments in standardization of high efficiency video coding (HEVC), Proc. SPIE, vol.7798, Aug. 2010.
[3]I. K. Kim, S. I. Sekiguchi, W. J. Han, K. McCann, and B. Bross, “High efficiency video coding (HEVC) test model 5 encoder description, Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11, Doc. JCTVC-G1102 Nov. 2011.
[4]S. Fukushima, H. Nakamura, and H. Takehara, “Simplification of AMVP spatial candidate derivation, Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11, Doc. JCTVC-H0078, February 2012.
[5]Y. Itani, S. Sekiguchi, and Y. Yamada, “Adaptive direct vector derivation for video coding, Proc. PCS, Dec. 2010, pp. 190-193.
[6]W. Yang, O. C. Au, C. Pang, J. Dai, F. Zou, X. Wen, and Y. Liu, “An efficient motion vector coding algorithm based on adaptive predictor selection, Proc. IEEE ISCAS, June 2010, pp. 2175-2178.
[7]Q. Yu, S. Ma, H. Liu, and J. Jia, “Parallel AMVP candidate list construction, Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11, Doc. JCTVC-I0036, May 2012.
[8]C. C. Lou, S. W. Lee and C. C. Jay Kuo, “Motion vector search window prediction in memory-constrained systems, Proc. SPIE, vol. 7443, pp. 1-12, Aug. 2009.
[9]C. C. Lou, S. W. Lee and C. C. Jay Kuo, “Adaptive search range selection in motion estimation, Proc. IEEE Speech Signal Process, Mar. 2010, pp. 918–921..
[10]Y. H. Ko, H. S. Kang, and S. W. Lee, “Adaptive search range motion estimation using neighboring MVDs, Proc. IEEE Transactions on Consumer Electronics, Vol. 57, No. 2, May 2011, pp. 726-730.
[11]C. C. Lou, S. W. Lee, and C. C. Jay Kuo, “Adaptive motion search range prediction for video encoding, Proc. IEEE Transactions on Circuits and Systems for Video Technology, vol. 20, Dec. 2010, pp. 1903-1908.
[12]Y. H. Ko, H. S. Kang, and S. W. Lee, “Adaptive search range motion estimation using neighboring motion vector differences, Proc. IEEE Transactions on Consumer Electronics, Vol. 57, No. 2, May 2011, pp. 726-730.
[13]S. Ryu, J. Seo, X. Liu, J. Y. Lee, H. Wey, and K. Sohn, “Analysis of motion vector predictor in multiview video coding system, Proc. IEEE International Symposium on Parallel and Distributed Processing with Applications Workshops, May 2011, pp. 184-188.
[14]S. C. Chang, J. F. Yang, and J. N. Hwang, , “Computation reduction algorithms of motion estimation and DCT Department of Electrical Engineering National Cheng Kung University Tainan, Taiwan, R.O.C. Dissertation for Doctor of Philosophy, May 2003.
[15]F. Bossen, “Common test conditions and software reference configurations, Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11, Doc. JCTVC-G1200, Nov. 2011.
[16]G. Bjontegaard, “Calculation of average PSNR differences between RD-curves, 13th VCEG-M33 Meeting: Austin, Texas, USA, 2-4 April, 2001.