Dicom Color Medical Image Compression using 3D-SPIHT for Pacs Application

Area of Science:

• Medical Imaging
• Image Compression
• Wavelet Transforms

Background:

• Volumetric medical imaging generates large datasets, necessitating efficient compression.
• Existing compression standards like JPEG 2000 may not fully exploit 3D spatial and temporal redundancies in medical image sequences.

Purpose of the Study:

• To apply and evaluate the 3D Set Partitioning In The Hierarchical Trees (3D-SPIHT) algorithm for color volumetric DICOM medical image compression.
• To assess the performance of 3D-SPIHT using biorthogonal 9/7 filters against JPEG 2000.

Main Methods:

• Utilized 3D wavelet decomposition with biorthogonal 9/7 filters for volumetric data.
• Applied a 3D spatial dependence tree and set partitioning techniques for progressive coding of wavelet coefficients.
• Implemented 3D-SPIHT and evaluated rate-distortion performance (PSNR vs. bit rate) on medical datasets.

Main Results:

• The 3D-SPIHT algorithm demonstrated superior performance compared to JPEG 2000 in terms of Peak Signal-to-Noise Ratio (PSNR) and compression ratio.
• The method effectively exploited color space relationships and inter-slice correlations for enhanced compression.
• Achieved full embeddedness suitable for color image sequence compression.

Conclusions:

• 3D-SPIHT offers a highly effective compression solution for color volumetric medical images.
• The algorithm's performance suggests its suitability for Picture Archiving and Communication Systems (PACS).
• This approach provides significant improvements in compression efficiency and image quality for medical imaging applications.