Improving 3D-TV view synthesis using motion compensated temporal interpolation

Nowadays, the development of three-dimension (3D) video applications such as three-dimensional television (3D-TV) and free-viewpoint television (FTV) has greatly increased human experiences. View synthesis method like depth-image-based-rendering (DIBR), plays a significant role in 3D content creation, 3D transmission, and has been integrated into video coding standards such as 3D-High efficiency video coding (3D-HEVC). However, the current DIBR method employs only the disparity correlation between views to create a so-called synthesized view; thus, unable to take full advantages of available synthesized information. In this paper, we propose a novel view synthesis method which takes advantages of not only the disparity correlation but also the temporal correlation between views. In the proposed method, an effective motion compensation based frame interpolation is employed to generate a temporal prediction view which is then combined with the DIBR rendered view to obtain the final synthesized view. Experimental results show that the proposed method can achieve the synthesized view with significantly outperforming other conventional techniques in terms of both peak signal-to noise ratio (PSNR) and subjective visual quality.