三維高效率視訊編碼(3D High Efficiency Video Coding, 3D-HEVC)是為了支援多視角影片(Multiview video)和基於深度資訊的3D影片(Depth-based 3D video)而設計的最新一代的三維視訊編碼標準。雖然能更有效地處理多視角加景深(Multiview plus depth, MVD)這種格式，但伴隨而來的是較高的編碼複雜度。本論文提出編碼單位深度預測、預測單位模式決策及位元失真成本(Rate-Distortion Cost, RD-cost)預測等快速演算法來減少3D-HEVC編碼複雜度。本論文提出透過空間域、時間域及視角間的對應編碼單位來預估當前編碼單位的深度及刪減不必要的預測單位模式，並且透過獨立視角中對應編碼單位的位元失真成本值來預測相依視角編碼單位之位元失真成本值，減少不必要的編碼程序。實驗結果顯示本論文所提出的演算法與原系統HTM-16.0比較，在Random-access架構下平均可節省約34.379%的編碼時間，而合成視角的平均BD-Bitrate只有增加1.757%，在相同的畫面品質下比參考文獻使用了更少的編碼資料量，並且能節省更多的編碼時間。

Three-Dimensional High Efficiency Video Coding (3D-HEVC), the extension of HEVC, is designed to support the multi-view and depth-based 3D video. By using 3D-HEVC, the multi-view plus depth (MVD) video format can be coded more efficiently. However, the improvement of video quality is accompanied by the increase of computation complexity.
In this thesis, we propose a fast algorithm with coding unit (CU) prediction, prediction unit (PU) mode decision and rate-distortion cost (RD-cost) prediction to solve the problem of high complexity of 3D-HEVC encoder. Our method uses the relationship between current CU and spatially, temporally, inter-view neighboring CUs to avoid unnecessary CU depths and PU modes. Early termination of the unnecessary CU process in dependent view is proposed by predicting the RD-cost of the CU in dependent view from the RD-costs of corresponding CUs in independent view.
The experimental results show that our algorithm can reduce 34.379% coding time on average, while the average BD-Bitrate of synthesized view is 1.757% compared to HTM16.0. Furthermore, our fast algorithm outperforms previous work for both coding performance and coding speed.

摘要 I
Abstract III
目錄 V
表目錄 VIII
圖目錄 IX
第一章 緒論 1
1.1 三維高效率視訊編碼概述 1
1.2 三維高效率視訊編碼系統架構介紹 2
1.2.1 獨立視角編碼 5
1.2.2 相依視角編碼 11
1.2.3 景深資訊編碼 15
1.3 研究動機 16
1.4 論文架構 16
第二章 相關文獻探討 17
2.1 應用於彩色影像之快速演算法 17
2.2 應用於景深資訊之快速演算法 18
2.3 應用於3D-HEVC整體之快速演算法 19
第三章 所提出之快速演算法 21
3.1 編碼單位深度預測與修正演算法 21
3.1.1 編碼單位深度預測 22
3.1.2 編碼單位深度修正 29
3.2 預測單位模式決策演算法 32
3.3 位元失真成本預測演算法 35
3.3.1 不同視角間位元失真成本分布之觀察 35
3.3.2 位元失真成本預測 35
3.4 快速編碼單位及預測單位決策演算法之整體流程 39
第四章 實驗結果 43
4.1 實驗環境及實驗參數之介紹 43
4.2 實驗結果 49
第五章 結論與未來展望 65
參考文獻 67

[1] Gary J. Sullivan, Jens-Rainer Ohm, Woo-Jin Han, and Thomas Wiegand, “Overview of the High Efficiency Video Coding (HEVC) Standard,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 22, no. 12, pp. 1649-1668, September 2012.
[2] Glenn Van Wallendael, Sebastian Van Leuven, Jan De Cock, Fons Bruls, and Rik Van de Walle, “3D Video Compression Based on High Efficiency Video Coding,” IEEE Transactions on Consumer Electronics, vol. 58, no. 1, pp. 137-145, February 2012.
[3] Gerhard Tech, Ying Chen, Karsten Müller, Jens-Rainer Ohm, Anthony Vetro, and Ye-Kui Wang, “Overview of the Multiview and 3D Extensions of High Efficiency Video Coding,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 26, no.1, pp. 35-49, January 2016.
[4] Ying Chen, Gerhard Tech, Krzysztof Wegner, and Sehoon Yea, “Test Model 11 of 3D-HEVC and MV-HEVC,” JCT3V-K1003, Geneva, February 2015.
[5] Chris Rosewarne, Benjamin Bross, Matteo Naccari, Karl Sharman, and Gray Sullivan, “High Efficiency Video Coding (HEVC) Test Model 16 (HM 16) Improved Encoder Description,” JCTVC-W1002, San Diego, February 2016.
[6] Antonio Ortega, and Kannan Ramchandran, “Rate-Distortion Methods for Image and Video Compression,” IEEE Signal Processing Magazine, vol. 15, no. 6, pp. 23-50, November 1998.
[7] Jewon Kang, Ying Chen, Li Zhang, and Marta Karczewicz, “3D-CE5.h related: Improvement for Disparity Vector Derivation,” JCT3V-B0047, Shanghai, October 2012.
[8] Yu-Lin Chang, Chi-Ling Wu, Yu-Pao Tsai, and Shawmin Lei, “CE1.h: Depth-Oriented Neighboring Block Disparity Vector (DoNBDV) with Virtual Depth Retrieval,” JCT3V-C0131, Geneva, January 2013.
[9] Jin Young Lee, Min Woo park, and Channel Kim, “3D-CE1: Depth Intra Skip (DIS) Mode,” JCT3V-K0033, Geneva, February 2015.
[10] Xu Chen, Xiaozhen Zheng, Sunmi Yoo, Sehoon Yea, Gun Bang, Young Su Heo, Woo When Gwun, Gwang Hoon Park, Gwang Soon Lee, and Nam Ho Hur, “Single Depth Intra Mode Simplification,” JCT3V-J0115, Strasbourg, October 2014.
[11] Kuang-Han Tai, Min-Yuan Hsieh, Mei-Juan Chen, Chia-Yen Chen, and Chia-Hung Yeh, “A Fast HEVC Encoding Method Using Depth Information of Collocated CUs and RD Cost Characteristics of PU Modes,” IEEE Transactions on Broadcasting, vol. 63, no. 4, pp. 680-692, July 2017.
[12] Chia-Hung Yeh, Ming-Feng Li, Mei-Juan Chen, Ming-Chieh Chi, Xin-Xian Huang, and Hao-wen Chi, “Fast Mode Decision Algorithm Through Inter-View Rate-Distortion Prediction for Multiview Video Coding System,” IEEE Transactions on Industrial Informatics, vol. 10, no. 1, pp. 594-603, July 2013.
[13] Na Zhang, Debin Zhao, Yi-Wen Chen, Jian-Liang Lin, and Wen Gao, “Fast Encoder Decision for Texture Coding in 3D-HEVC,” Signal Processing: Image Communication, vol. 29, no. 9, pp. 951-961, October 2014.
[14] Na Zhang, Yi-Wen Chen, Jian-Liang Lin, Jicheng An, Kai Zhang, Shawmin Lei, Siwei Ma, Debin Zhao, and Wen Gao, “3D-CE3.h related: Fast Encoder Decision for Texture Coding,” JCT3V-E0173, Vienna, July 2013.
[15] Yubing Wang, Yongfang Wang, and Yawen Shi, “A Fast CU Size Decision Algorithm for 3D-HEVC,” MATEC Web of Conferences The International Seminar on Applied Physics, Optoelectronics and Photonics (APOP), vol. 61, June 2016.
[16] J. Jung, and K.Viswanathan, “Depth Quadtree Prediction for HTM,” JCT3V-A0044, Stockholm, July 2012.
[17] J. Jung, and E. Mora, “3D-CE3.h: Depth Quadtree Prediction for 3DHTM 4.1,” JCT3V-B0068, Shanghai, October 2012.
[18] Xianguo Zhan, Kai Zhang, Jicheng An, Jian-Liang Lin ,and Shawmin Lei, “3D-CE2 related: A Texture-Partition-Dependent Depth Partition for 3D-HEVC,” JCT3V-G0055, San Jose, January 2014.
[19] Ruhan Conceição, Giovanni Avila, Guilherme Corrêa, Marcelo Porto, Bruno Zatt, and Luciano Agostini, “Complexity Reduction for 3D-HEVC Depth Map Coding Based on Early Skip and Early DIS Scheme,” IEEE International Conference on Image Processing (ICIP), pp. 1116-1120, Phoenix, USA, September 2016.
[20] Ming Chen, Yongshuang Yang, Qiuwen Zhang, Xiaoxin Zhao, Xinpeng Huang, and Yong Gan, “Low Complexity Depth Mode Decision for HEVC-Based 3D Video Coding,” Optik - International Journal for Light and Electron Optics, vol. 127, no. 11, pp. 4758-4767, June 2016.
[21] Qiuwen Zhang, Na Zhang, Tao Wei, Kunqiang Huang, Xiaoliang Qian, and Yong Gan, “Fast Depth Map Mode Decision Based on Depth-Texture Correlation and Edge Classification for 3D-HEVC,” Journal of Visual Communication and Image Representation, vol. 45, pp. 170-180, May 2017.
[22] Qiuwen Zhang, Zhifeng Zhang, Bin Jiang, Xiaoxin Zhao, and Yong Gan, “Fast 3D-HEVC Encoder Algorithm for Multiview Video Plus Depth Coding,” Optik - International Journal for Light and Electron Optics, vol. 127, no. 20, pp. 8864-8873, October 2016.
[23] Pin-Chen Kuo, Kuan-Hsing Lu, Yun-Ning Hsu, Bin-Da Liu, and Jar-Ferr Yang, “Fast Three-Dimensional Video Coding Encoding Algorithm Based on Edge Information of Depth Map,” IET Image Processing, vol. 9, no. 7, pp. 587-595, July 2015.
[24] 3D-HEVC reference software version 16.0 (HTM-16.0), available online at https://hevc.hhi.fraunhofer.de/svn/svn_3DVCSoftware/tags/HTM-16.0/, accessed on January 2018.
[25] Karsten Müller, and Anthony Vetro, “Common Test Conditions of 3DV Core Experiments,” JCT3V-G1100, San José, January 2014.
[26] Gisle Bjøntegaard, “Calculation of Average PSNR Differences Between RD Curves,” ITU-T SG16/Q6, Document VCEG-M33, Austin, April 2001.
[27] Gisle Bjøntegaard, “Improvement of the BD-PSNR Model,” ITU-T SG16/Q6, Document VCEG-AI11, Berlin, July 2008.