Multiview plus depth scalable coding in the depth domain

Abstract

Three dimensional (3D) TV is a growing area that provides an extra dimension at the cost of spatial resolution. The multiview plus depth representation provides a lower bit rate when it is encoded than multi-view and higher resolution than a 2D-plus-depth sequence. Scalable video coding provides adaption to the conditions at the receiver. In this paper we propose a scheme that combines scalability in both the view and depth domain. The center view data is preserved, whereas the data of the side views are extracted in layers depending on distance to the camera. This allows a decrease in bit rate of 16-39 % for the colour part of a 3-view MV depending number of pixels in the first enhancement layer if one layer is extracted. Each additional layer increases the visual quality and PSNR compared only using center view data.

Layer Assignment Based on Depth Data Distribution for Multiview-Plus-Depth Scalable Video Coding

“…The present work is an extension to authors' previous works [20] and [21], where an approach depth and view scalability of a MVD sequence of three views was proposed. The depth scalability concerns scalability of the color data in relation to the distance to the camera.…”