Multi-view video compression using dynamic background frame and 3D motion estimation

摘要

The H.264/MVC multi-view video coding standard provides a better compression rate compared to the simulcast coding technique (i.e., H.264/AVC) by exploiting inter- and intraview redundancy. However, this technique imposes random access frame delay as well as requires huge computational time. In this paper three novel techniques are proposed to overcome the above mentioned problems. Firstly, a simulcast video coding technique is proposed where each view is encoded individually using two reference frames-immediate previous frame and a dynamic background frame (popularly known as McFIS- the most common frame in a scene) of the corresponding view. Secondly a novel technique is proposed using 3D motion estimation (3D-ME) where a 3D frame is formed using the same temporal frames of all views and ME is carried out for the current 3D frame using the immediate previous 3D frame as a reference frame. Thereafter, a fractional ME refinement is also conducted on individual frames of 3D current frame using individual reference frames. Finally, a modification of the 3D-ME technique is proposed where an extra reference frame namely 3D McFIS is used for 3D-ME. As the correlation among the intra-view images is higher compared to the correlation among the inter-view images, the proposed 3D-ME techniques reduce the overall computational time and eliminate the frame delay with comparable rate-distortion (RD) performance compared to H.264/MVC. Experimental results reveal that the proposed techniques outperform the H.264/MVC in terms of improved RD performance by reducing computational time and by eliminating the random access frame delay.