Displaying radiologic images on personal computers: image storage and compression--Part 2

Area of Science:

• Medical Imaging
• Computer Science
• Image Processing

Background:

• Radiologic image display, manipulation, and analysis on personal computers require efficient storage and compression.
• Image compression is crucial for managing large datasets in medical imaging.
• Understanding different compression algorithms is essential for radiologists and imaging scientists.

Purpose of the Study:

• To evaluate and compare lossless and lossy image compression techniques for radiologic images.
• To identify optimal compression methods for enhancing radiologic image storage and transmission.
• To analyze the effectiveness of various algorithms, including Huffman codes, Lempel-Ziv-Welch, and differential pulse-code modulation.

Main Methods:

• Classification of image compression into lossless (nondestructive) and lossy (destructive) categories.
• Implementation and testing of common lossless algorithms: variable-length bit codes (Huffman), dictionary-based (Lempel-Ziv variants, LZW), and arithmetic coding.
• Application of differential pulse-code modulation (DPCM) transformation prior to compression.
• Exploration of lossy compression steps: transformation (discrete cosine, discrete wavelet), quantization, and coding.
• Investigation of fractal-based image compression.

Main Results:

• Lossless compression methods like Huffman codes and Lempel-Ziv-Welch (LZW) are effective for radiologic images.
• Transforming images using differential pulse-code modulation (DPCM) enhances the performance of lossless algorithms.
• LZW compression following DPCM transformation yielded the best lossless results, comparable to commercial software.
• Lossy compression achieves higher data reduction but may introduce noticeable quality degradation and artifacts.
• Lossy compression techniques involve transformation, quantization, and coding to reduce data size.

Conclusions:

• Differential pulse-code modulation combined with Lempel-Ziv-Welch offers superior lossless compression for radiologic images.
• Lossy compression methods provide significant data reduction but require careful consideration of image quality trade-offs.
• The choice between lossless and lossy compression depends on the specific application and acceptable quality levels for radiologic analysis.