A zerotree stereo video encoder

Abstract

Over the past decade, the growing interests in realistic visual communication systems have led to the rapid development of 3D visual compression techniques. A stereoscopic system is one way of simulating 3D perception. Without exploiting the high redundancy in a stereo pair, the amount of data required to store or transmit it is doubled, when compared to that for monocular images. In this paper, we present a new technique for coding stereo video sequences based on the Discrete Wavelet Transform (DWT). The proposed technique exploits Zerotree Entropy Coding (ZTE) that makes use of the wavelet block concept to achieve low bit rate stereo video coding. One of the two image streams, the main stream, is independently coded by a modified zerotree video coder and the second stream, the auxiliary stream, is predicted based on disparity compensation. The residual image of the auxiliary stream is then coded by a modified zerotree video coder. Results show that the proposed encoder can achieve 50% bandwidth reduction compared to the ZTE based monoscopic compression encoder.