Efficient implementation and evaluation of wavelet packet for 3D medical image segmentation

3D volume segmentation aims at partitioning the voxels into 3D objects (sub-volumes) which represent meaningful physical entities. Multi-resolution analysis (MRA) allows for the preservation of an image according to certain levels of resolution or blurring. The quality of this approach makes it useful in image compression, de-noising, and classification or segmentation. This paper focuses on the implementation of a medical volume segmentation technique using different wavelet decompositions such as discrete wavelet transform (DWT) and discrete wavelet packet transform (WP). A comparison study has been carried out to evaluate both decompositions using different performance metrics such as Euclidean distance (ED) and dice similarity coefficients (DSC) which reveal that WP approaches can accurately detect the perform the details coefficients in both phantom and real data.