基于高效率视频编码(high efficiency video coding, HEVC)三维视频扩展的深度编码,引入了深度模型编码模式(depth modeling mode, DMM)和区域边界链编码(region boundary chain, RBC)模式. RBC 模式通过遍历编码单元(coding unit, CU)当前深度下所有的边缘线预测模板来得到最佳的预测模板, 但在显著提升编码效率的同时, 其计算复杂度也增加了数倍. 深入研究了在HEVC 模式粗选过程中选出的最优预测模式与DMM 和RBC 这两种模式的相关性, 以及与直接Wedgelet 搜索模板集合各个方向的相关性, 并在此基础上提出了一种快速的自适应深度图帧内预测模式选择算法. 该算法将编码块分为平坦块、方向性编码块以及纹理复杂块3 种. 对平坦块跳过DMM 和RBC 模式, 而对方向性编码块的直接Wedgelet 搜索过程则跳过不必要的方向模板搜索, 从而提高编码速度. 实验结果表明, 在全I 帧模式下,该深度图帧内预测模式选择方法平均节省了75.4% 的编码时间, 而在合成视点端仅带来0.4%的性能损失.
In the high efficiency video code (HEVC) based depth video coding in 3D video coding, depth modeling mode (DMM) and a region boundary chain (RBC) mode
are introduced. DMM traverses all Wedgelet patterns in the current coding unit (CU) to get the best prediction pattern. Encoding efficiency is improved significantly by these added technologies. Computation complexity is increased significantly, hindering its practical application. To reduce computation complexity, a fast depth mode decision algorithm is proposed for depth video coding based on correlation between the most probable mode and DMM and RBC mode. The proposed method divides coding blocks into flat blocks, directional blocks and texture-complicated blocks. Redundant DMM and RBC mode are
skipped for flat blocks. Explicit Wedgelet search patterns are skipped for directional blocks based on directivity of non-flat blocks. Experimental results show that the proposed algorithm can reduce computational complexity of depth coding by 75.4% for all-intra test case, with BD-rate increase of only 0.4% after decoder-side view synthesis.
[1] Rusanovskyy D, Chen F C, Zhang L. 3D-AVC test model 7 [C]// ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 11 8th Meeting, JCT3V. 2013: E1003.
[2] Zhang L, Tech G, Wegner K, et al. 3D-HEVC test model 5 [C]// ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 11 8th Meeting, JCT3V. 2013: E1005.
[3] 胡锦雯, 滕国伟, 成益龙, 等. 基于运动特征的HEVC 快速帧间预测算法[J]. 上海大学学报: 自然科学版, 2013, 19(3): 245-249.
[4] Jager F, Ball´e J. Median trilateral loop filter for depth map video coding [C]// Picture Coding Symposium (PCS). 2012: 61-64.
[5] Liu S, Lai P L, Tian D, et al. New depth coding techniques with utilization of corresponding video [J]. IEEE Transactions on Broadcasting, 2011, 57(2): 551-561.
[6] Merkle P, Morvan Y, Smolic A, et al. The effects of multiview depth video compression on multiview rendering [J]. Signal Processing: Image Communication, 2009, 24(1): 73-88.
[7] Fehn C. Depth-image-based rendering (DIBR), compression, and transmission for a new approach on 3D-TV [C]// Electronic Imaging 2004, International Society for Optics and
Photonics. 2004: 93-104.
[8] Oh B, Lee J, Park D. Depth map coding based on synthesized view distortion function [J]. IEEE Journal of Selected Topics in Signal Processing, 2011, 5(7): 1344-1352.
[9] Philipp M, Karsten M. CE5.H: results on DMM and RBC depth intra modes performance (Test 1) [C]// ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 11 8th Meeting, JCT3V.
2013: E0120.
[10] Gu Z Y, Zheng J H, Ling N, et al. 3D-CE5.H related: fast intra prediction mode selection for intra depth map coding [C]// ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 11 8th
Meeting, JCT3V. 2013: E0238.
[11] Zhang M, Zhao C, Xu J, et al. A fast depth-map wedgelet partitioning scheme for intra prediction in 3D video coding [C]// IEEE International Symposium on Circuits and Systems.
2013: 2852-2855.
[12] M¨uller K, Merkle P, Tech G, et al. 3D video coding with depth modeling modes and view synthesis optimization [C]// Signal & Information Processing Association Annual Summit and
Conference. 2012: 1-4.
[13] Han G J, Ohm J R, Han W J, et al. Overview of the high efficiency video coding (HEVC) standard [J]. IEEE Transactions on Circuits and Systems for Video Technology, 2012, 22(12):
1649-1668.
[14] Bjontegaard G. Improvements of the BD-PSNR model [C]// 35th VCEG Meeting. 2008: 16-18.
[15] Rusanovskyy D, M¨uller K, Vetro A. Common test conditions of 3DV core experiments [C]// ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 11 8th Meeting, JCT3V. 2013: E1100.
