| Region-based coding is an important feature provided in today's image coding schemes including SPIHT and JPEG2000 as it allows different regions of interest in an image to be encoded at different bit rates and hence at different qualities rather than encoding the entire image with a single quality constraint. In this thesis, we consider three-dimensional region-based coding of volumetric medical datasets and video sequences. A 3D SA-DWT is used to decompose the data with multiple, arbitrarily shaped regions to obtain the representation of the regions in the transform domain. Then, a modified 3D SPIHT coding algorithm based on an unbalanced tree structure is adopted in 3D region-based coding. This coding scheme offers good rate-distortion performance with additional features such as distortion scalability and flexibility in precise rate control. Experimental results show that the proposed algorithm outperforms the other coding schemes based on SPIHT algorithm in terms of R-D performance. |
