Lossless Compression of Volumetric Medical Images with Improved 3-D SPIHT Algorithm
Sungdae Cho, Dongyoun Kim, William A. Pearlman · 2003
This paper presents a lossless compression of volumetric medical images with the improved 3-D SPIHT algorithm that searches on asymmetric trees. The tree structure links wavelet coefficients produced by three-dimensional reversible integer wavelet transforms. Experiments show that the lossless compression with the improved 3-D SPIHT gives improvement about 42% on average over two-dimensional techniques, and is superior to those of prior results of three-dimensional techniques. In addition to that, we can easily apply different numbers of decompositions between the transaxial and axial dimensions, which is a desirable function when the coding unit of a group of slices is limited in size.