A Full Scalable Volumetric Medical Image Coding for Efficient Archiving and Transmission Applications

A fully scalable 3D wavelet coding approach for lossy-to-lossless coding of volumetric medical im age data is proposed. The proposed approach, called 3DHS-SPIHT, divides the 3D input medical images into groups of slices (GOS) and encodes each GOS as a separate unit. The symmetric tree definition of the original 3DSPIHT is improved by introducing an asymmetric tree structure. While preserving com pression efficiency, the new tree structure allows for a small size for each GOS which reduces memory consumption and also facilitates more efficient random access to the slices. The experimental results on MR data sets show the efficiency of the 3DHS-SPIHT algorithm for multiresolution lossy-to-lossless cod ing, which make it a very attractive coding candidate for volumetric medical image information archiving and transmission applications.

Keywords:

Medical image coding , volumetric coding , 3DSPIHT , scalability