Int J Performability Eng ›› 2018, Vol. 14 ›› Issue (9): 1968-1974.doi: 10.23940/ijpe.18.09.p4.19681974

Previous Articles     Next Articles

Fast 3D-HEVC Coding based on Support Vector Machine

Hanqing Dinga, *, Shuaichao Weia, Yan Zhangb, and Qiuwen Zhanga   

  1. aCollege of Computer and Communication Engineering, Zhengzhou University of Light Industry, Zhengzhou, 450002, China;
    bSchool of Computer and Remote Sensing Information Technology, North China Institute of Aerospace Engineering, Langfang, 065000, China
  • Revised on ; Accepted on
  • Contact: * E-mail address: hanqingding2013@163.com

Abstract: 3D high efficiency video coding (3D-HEVC) is the latest video compression standard for multi-view video systems. In this paper, a fast coding method is proposed by utilizing machine learning to alleviate the complexity of the 3D-HEVC system while maintaining the RD performance. The main content of our algorithm is to utilize the support vector machine (SVM) to analyze the motion properties of texture video where variable mode prediction is needed and early skip unnecessary modes for a given coding unit (CU). Experimental results confirmed that the proposed method could greatly lessen the computational complexity of the 3D-HEVC system with only a small BD-rate loss for texture view and synthesized view.

Key words: 3D-HEVC system, motion properties, SVM