Weighted Large Margin Nearest Center Distance-Based Human Depth Recovery With Limited Bandwidth Consumption

doi:10.1109/TIP.2018.2855414

	Weighted Large Margin Nearest Center Distance-Based Human Depth Recovery With Limited Bandwidth Consumption
	Huang, Meiyu 1; Xiang, Xueshuang 1; Chen, Yiqiang 2,3; Fan, Da 1
	2018-12-01
发表期刊	IEEE TRANSACTIONS ON IMAGE PROCESSING
ISSN	1057-7149
卷号	27 期号:12 页码:5728-5743
摘要	This paper proposes a weighted large margin nearest center (WLMNC) distance-based human depth recovery method for tele-immersive video interaction systems with limited bandwidth consumption. In the remote stage, the proposed method highly compresses the depth data of the remote human into skeletal block structures by learning the WLMNC distance, which is equivalent to downsampling the human depth map at 64x the sampling rate. In the local stage, the method first recovers a rough human depth map based on a WLMNC distance augmented clustering approach and then obtains a fine depth map based on a rough depth-guided autoregressive model to preserve the depth discontinuities and suppress texture copy artifacts. The proposed WLMNC distance is learned by the large margin clustering problem with a weighted hinge loss to balance the clustering accuracy and depth recovery accuracy and is verified to be able to preserve depth discontinuities between skeletal block structures with occlusion. A theoretical analysis is conducted to verify the effectiveness of using the weighted hinge loss. Furthermore, a novel data set containing various types of human postures with self-occlusion is built to benchmark the human depth recovery methods. The quantitative comparison with the state-of-the-art depth recovery methods on the introduced benchmark data set demonstrates the effectiveness of the proposed method for human depth recovery with such a high upsampling rate.
关键词	Tele-immersive interaction occlusion handling depth recovery distance learning
DOI	10.1109/TIP.2018.2855414
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[61702520] ; National Natural Science Foundation of China[61773383] ; National Natural Science Foundation of China[61572471] ; National Key Research and Development Program of China[2017YFC0803401] ; Beijing Municipal Science and Technology Commission[Z171100000117017] ; Innovation Foundation of Qian Xuesen Laboratory of Space Technology
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Artificial Intelligence ; Engineering, Electrical & Electronic
WOS记录号	WOS:000444019600001
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计	被引频次：4[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/4946
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Xiang, Xueshuang
作者单位	1.China Acad Space Technol, Qian Xuesen Lab Space Technol, Beijing 100094, Peoples R China 2.Chinese Acad Sci, Beijing Key Lab Mobile Comp & Pervas Device, Inst Comp Technol, Beijing 100190, Peoples R China 3.Univ Chinese Acad Sci, Sch Comp & Control Engn, Beijing 100049, Peoples R China
推荐引用方式 GB/T 7714	Huang, Meiyu,Xiang, Xueshuang,Chen, Yiqiang,et al. Weighted Large Margin Nearest Center Distance-Based Human Depth Recovery With Limited Bandwidth Consumption[J]. IEEE TRANSACTIONS ON IMAGE PROCESSING,2018,27(12):5728-5743.
APA	Huang, Meiyu,Xiang, Xueshuang,Chen, Yiqiang,&Fan, Da.(2018).Weighted Large Margin Nearest Center Distance-Based Human Depth Recovery With Limited Bandwidth Consumption.IEEE TRANSACTIONS ON IMAGE PROCESSING,27(12),5728-5743.
MLA	Huang, Meiyu,et al."Weighted Large Margin Nearest Center Distance-Based Human Depth Recovery With Limited Bandwidth Consumption".IEEE TRANSACTIONS ON IMAGE PROCESSING 27.12(2018):5728-5743.