Int J Performability Eng ›› 2019, Vol. 15 ›› Issue (4): 1094-1102.doi: 10.23940/ijpe.19.04.p4.10941102

Previous Articles     Next Articles

Efficient Inter-View Prediction Structure for Multi-View High Efficiency Video Coding

Tao Yana, *, In-Ho Rab, *, and Qian Zhangc   

  1. a School of Information Engineering, Putian University, Putian, 351100, China;
    b School of Computer, Information and Communication Engineering, Kunsan National University, Gunsan, 54150, South Korea;
    c School of Information and Electromechanical Engineering, Shanghai Normal University, Shanghai, 200234, China
  • Revised on ; Accepted on
  • Contact: E-mail address: yantaoshu@aliyun.com
  • About author:Tao Yan received his Ph.D. in communication and information systems from Shanghai University in 2010. He is an associate professor in the School of Information Engineering at Putian University. His major research interests include Multiview high efficiency video coding, rate control, and video codec optimization. In-Ho Ra received his Ph.D. in computer engineering from Chung-Ang University in 1995. He is a professor in the School of Computer, Information and Communication at Kusan National University. His major research interests include blockchain. Qian Zhang received her Ph.D. in communication and information systems from Shanghai University in 2011. She is an associate professor in the School of Information and Electromechanical Engineering at Shanghai Normal University. Her major research interests include Multiview high efficiency video coding.

Abstract: In addition to high coding efficiency, multiview high-efficiency video coding (MV-HEVC) should include backward compatibility and temporal random access, which are mainly determined by the prediction structure used. For standardization of multiview video coding, (MVC) MV-HEVC normally uses the fixed-view temporal prediction structure, which runs into challenges with various characteristics of multiview videos. This paper starts by exploring the relationship between MV-HEVC prediction structure and coding complexity, compression efficiency, and random access performance and constructs mathematical models; it then comprehensively considers coding efficiency and user random access based on multiview video similarity analysis and adaptively adjusts the inter-view prediction structure to obtain better coding performance. The experimental results demonstrate that when compared with MV-HEVC, the proposed method has better random access performance and improved coding efficiency.

Key words: multi-view high efficiency video coding, prediction structure, random access, correlation analysis