3D video coding based on mixed transforms

Abstract

In this paper a simple, yet efficient variation of 3D video coding has been proposed. The algorithm uses both DCT and DWT to exploit the temporal and spatial redundancies present in the video. A new tree structure has been devised. Performance comparisons of this coder is made with existing 3D wavelet coder using 3D-SPIHT algorithm. Results obtained are comparable to the existing coder. No motion compensation is used. This coder also has the interesting feature of scalability (spatial and SNR) without which its difficult to sustain in the world of Internet and mobile communication. Packetisation is carried out using packetized zero tree wavelet (PZW) using the proposed new tree structure which is resilient to packet erasures.