The wavelet transform is a very effective method for compressing a 3D medical image data set yielding a high compression ratio image with good quality. Figure 5 shows the block diagrams of 3D wavelet transform compression and decompression. In the compression process, a 3D wavelet transform is first applied to the 3D image data set, resulting in a 3D multiresolution representation of the image. Then the wavelet coefficients are quantized using scalar quantization. Finally, run-length and
Huffman coding are used to impose entropy coding on the quantized data. These steps are described in Section 4.
The decompression process is the inverse of the compression process. The compressed data are first entropy decoded, a dequantization procedure is applied to the decoded data, and the inverse 3D wavelet transform is used, resulting in the reconstructed 3D image data.
Was this article helpful?