Scalable multi-resolution video coding using a combined subband-DCT approach

A combined subband-DCT approach for scalable multi-resolution video coding is presented. Therefore, each image is decomposed into 4 subbands by analysis filtering, whereat the low frequency subband represents the base layer, and together with the high frequency subbands the enhancement layer is reconstructed by synthesis filtering. The low frequency subband and the three high frequency subbands are separately coded using DCT-based techniques similar to MPEG-2. Further, for adjusting the bit rate distribution between base and enhancement layer, SNR scalability for the base layer is used. Compared with non-scalable MPEG-2 coding, the additional overhead for scalability is only 5-9% of the total bit rate, whereas the MPEG-2 Spatial Scalable Profile (SSP) leads to an overhead of 66-70%. 1. Introduction Scalable multi-resolution video coding requires that parts of the encoded and transmitted data can be decoded separately by a base layer decoder to reconstruct low resolution images. In com..