一种深度图的时域下采样编码及重建方法

doi:10.3969/j.issn.1001-2400.2016.04.028

西安电子科技大学学报 ›› 2016, Vol. 43 ›› Issue (4): 160-165+171.doi: 10.3969/j.issn.1001-2400.2016.04.028

一种深度图的时域下采样编码及重建方法

葛川1;刘琚1,2;元辉1,3;肖依凡1;李凤荣3

(1. 山东大学信息科学与工程学院，山东济南 250101;
2. 山东大学苏州研究院，江苏苏州 215123;
3. 中国科学院上海微系统与信息技术研究所无线传感网与通信重点实验室，上海 200050)

收稿日期:2015-06-29 出版日期:2016-08-20 发布日期:2016-10-12
通讯作者: 刘琚(1965-)，男，教授，博士，E-mail：juliu@sdu.edu.cn
作者简介:葛川(1985-)，男，山东大学博士研究生，E-mail：gavinkeh.chuan@gmail.com．
基金资助:
国家自然科学基金资助项目(61201211, 61571274)；教育部博士点基金资助项目(20120131120032，20130131110029)；山东省优秀中青年科学家奖励基金资助项目(BS2012DX021)；中国科学院无线传感网与通信重点实验室开放课题资助项目(2013002)；山东大学青年学者未来计划资助项目(2015WLJH39)；苏州市科技计划资助项目(SYG201443)

Temporal subsampling based depth maps coding and the reconstruction method

GE Chuan1;LIU Ju1,2;YUAN Hui1,3;XIAO Yifan1;LI Fengrong3

(1. School of Information Science and Engineering, Shandong Univ., Jinan 250101, China;
2. Suzhou Research Institute of Shandong Univ., Suzhou 215123, China;
3. Key Lab. of Wireless Sensor Network ＆ Communication, Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences, Shanghai 200050, China)

Received:2015-06-29 Online:2016-08-20 Published:2016-10-12

摘要/Abstract

摘要：

为提高三维视频系统的编码效率，提出了一种针对中间视点的深度时域下采样编码方法．首先确定深度图像的丢弃方式；然后利用基于时间一致性和视点间的相关性，恢复出丢弃的深度图像，并对恢复出的结果进行维纳滤波，进一步提高恢复的深度图的质量．该方法需要向编码端发送维纳滤波器系数和重建深度选择标识符．实验结果表明，在保证同等码率的前提下，解码端重建的虚拟视图的峰值信噪比最大能提升0.229dB，平均提升0.130dB，编码性能得到改善．

关键词: 视频编码, 立体图像处理, 深度图时域下采样, 时间一致性, 视点间相关性

Abstract:

In order to improve the coding efficiency of a three-dimensional video system, a depth temporal subsampling based coding method for the intermediate view is proposed. In this paper, the manner of how to discard depth frames is firstly determined. Then, those discarded depths are reconstructed utilizing temporal consistency and multiview correspondences. Finally, in order to further improve the quality of the reconstructed depth, the wiener filter is implemented on the reconstructed results. Filter coefficients and indicators for selecting the depth predictor should be transmitted to the decoder. Experimental results demonstrate that a maximum 0.229dB and average 0.130dB peak signal-to noise ratio (PSNR) gain could be achieved for the virtual view reconstructed by the decoder while maintaining the same coding bit rate, and that the coding efficiency can be improved.

Key words: video coding, stereo image processing, depth temporal subsampling, temporal consistency, multiview correspondences

葛川;刘琚;元辉;肖依凡;李凤荣;. 一种深度图的时域下采样编码及重建方法[J]. 西安电子科技大学学报, 2016, 43(4): 160-165+171.

GE Chuan;LIU Ju;YUAN Hui;XIAO Yifan;LI Fengrong. Temporal subsampling based depth maps coding and the reconstruction method[J]. Journal of Xidian University, 2016, 43(4): 160-165+171.

参考文献

［1］ GE C, LIU J, YUAN H. Optimizing Rate Allocation between Multiview Videos and Associated Depth Maps with Quantization-based Virtual View Distortion Model and Genetic Algorithm ［J］. Journal of Electronic Imaging, 2014, 23(6): 063016.
［2］马祥, 霍俊彦, 任光亮, 等. 利用视频和深度图相关性的深度图帧内编码［J］. 西安电子科技大学学报, 2015, 42(3): 1-7.
MA Xiang, HUO Junyan, REN Guangliang, et al. Depth Map Intra Coding Based on Correlation Between Video and Depth Maps ［J］. Journal of Xidian University, 2015, 42(3): 1-7.
［3］ FEHN C. Depth-image-based Rendering (DIBR), Compression, and Transmission for a New Approach on 3D-TV ［C］//Proceedings of SPIE: 5291. Bellingham: SPIE, 2004: 93-104.
［4］ YUAN H, KWONG S, GE C, et al. Interview Rate Distortion Analysis-Based Coarse to Fine Bit Allocation Algorithm for 3-D Video Coding［J］. IEEE Transactions on Broadcasting, 2014, 60(4): 614-625.
［5］ YUAN H, KWONG S, LIU J, et al. A Novel Distortion Model and Lagrangian Multiplier for Depth Maps Coding ［J］. IEEE Transactions on Circuits and Systems for Video Technology, 2014, 24(3): 443-451.
［6］ YUAN H, LIU J, LI Z, et al. Virtual View Oriented Distortion Criterion for Depth Maps Coding ［J］. Electronics Letters, 2012, 48(1): 23-25.
［7］ GRAZIOSI D B, RODRIGUES N M M, PAGLIARI C L, et al. Compressing Depth Maps using Multiscale Recurrent Pattern Image Coding［J］. Electronics Letters, 2010, 46(5): 340-341.
［8］ OH K, YEA S, VETRO A, et al. Depth Reconstruction Filter and Down/Up Sampling for Depth Coding in 3-D Video［J］. IEEE Signal Processing Letters, 2009, 16(9):747-750.
［9］ EKMEKCIOGLU E, WORRALL S T, KONDOZ A M. A Temporal Subsampling Approach for Multiview Depth Map Compression［J］. IEEE Transactions on Circuits and Systems for Video Technology, 2009, 19(8): 1209-1213.
［10］ LIU C. Beyond Pixels: Exploring New Representations and Applications for Motion Analysis ［D］. Cambridge: Massachusetts Institute of Technology, 2009.
［11］ YUAN H, LIU J, XU H, et al. Coding Distortion Elimination of Virtual View Synthesis for 3D Video System: Theoretical Analyses and Implementation ［J］. IEEE Transactions on Broadcasting, 2012, 58(4): 558-568.
［12］ JOINT COLLABORATIVE TEAM FOR 3DV. 3D-HTM Software Platform［EB/OL］. ［2015-06-13］. https://hevc.hhifraunhofer.de/svn/svn_3DVCSoftware/tags/.
［13］ BJONTEGAARD G. Calculation of Average PSNR Differences between RD-curves ［C］//Processings of 13th Meeting of the ITU-T Video Coding Experts Group. Austin: ACEG, 2001:1-4.

[1]	李强,左静,王海宁. SHVC中帧内预测快速算法[J]. 西安电子科技大学学报, 2020, 47(2): 60-66.
[2]	吴炜;陈健;宋彬. 一种新的H.264/AVC视频编码失真估计算法[J]. J4, 2013, 40(3): 1-6+35.
[3]	蒋洁;郭宝龙;莫玮;焦昶哲;孙伟. 利用平滑区域检测的HEVC帧内编码快速算法[J]. J4, 2013, 40(3): 194-200.
[4]	秦浩;屈蓓;宋彬;杨明明. 无反馈分布式视频编码中Wyner-Ziv帧码率控制算法[J]. J4, 2012, 39(4): 33-38+93.
[5]	肖嵩;刘婷;杜建超;卢冀. 分级视频码流的三维精细粒度排序方法[J]. J4, 2012, 39(2): 73-79.
[6]	元辉;常义林;卢朝阳;李明. 一种降低预测模式开销的帧内预测方法[J]. J4, 2010, 37(6): 981-986+1016.
[7]	李明;常义林;杨付正;万帅. 一种自适应的关键帧量化参数确定方法[J]. J4, 2010, 37(3): 396-404+435.
[8]	卢冀;肖嵩;吴成柯. 基于网络编码的SVC高效传输系统[J]. J4, 2010, 37(3): 405-411.
[9]	邓云1;2;彭强1;诸昌钤1. 一种自适应的AR-FGS漏因子选择算法 [J]. J4, 2008, 35(6): 1115-1120.
[10]	杨付正(1);常义林(1);王新岱(1);王静(2). 基于水印的视频误码掩盖方法[J]. J4, 2005, 32(1): 16-20.
[11]	杨清永;王新;周俊辉. 块运动估计中具有可扩展性的高精度三步搜索算法[J]. J4, 2002, 29(5): 610-614.
[12]	卢朝阳;周幸妮;高西全. MPEG标准中的三角形网格基运动补偿方法[J]. J4, 2001, 28(6): 816-820.
[13]	宋彬;常义林;郭建英. 基于H.263的甚低比特率视频软件编码器[J]. J4, 2000, 27(1): 52-58.
[14]	常义林;孙献璞;吴剑. 视频变速率编码的统计模型研究[J]. J4, 1996, 23(2): 0-0.

一种深度图的时域下采样编码及重建方法

Temporal subsampling based depth maps coding and the reconstruction method

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 14

Metrics

本文评价

推荐阅读 10