Abstract
In this paper, the potential for improving the compression efficiency of multi-view video coding with depth information is explored. The proposed technique uses downsampling prior to encoding, for arbitrary views and depth maps. A bit-rate adaptive downscaling-ratio decision approach is proposed for certain views and depth maps prior to encoding. Colour and depth videos are considered separately due to their different characteristics and effects on synthesized free view-point videos. The inter-view references, if present, are downsampled to the same resolution as the input video to be coded. The results for several multi-view with depth sequences indicate that using bit-rate adaptive mixed spatial resolution coding for both views and depth maps can achieve savings in bit-rate, compared to full resolution and fixed depth-to-colour ratio multi-view coding when the quality of synthesized viewpoints are considered. The computational complexity in the encoder is significantly reduced at the same time, since the number of blocks coded is reduced, and hence the number of block mode decisions carried out is reduced.