Quadtree-based data compression of abdominal CT images.

Area of Science:

• Medical Imaging
• Computer Science
• Radiology

Background:

• Diagnostic imaging relies on high-quality image data.
• Data compression is crucial for efficient storage and transmission of medical images.
• Quadtree-based compression is a technique explored for its efficiency.

Purpose of the Study:

• To evaluate the impact of quadtree-based data compression on the diagnostic accuracy of computed tomography (CT) images.
• To assess the trade-offs between compression ratio and diagnostic yield in abdominal/pelvic CT scans.

Main Methods:

• Receiver Operating Characteristic (ROC) study methodology was employed.
• Seventy abdominal/pelvic CT images were analyzed.
• Images were compressed using varying irreversible ratios (30.6:1 to 7.4:1) and reversible compression (2.9:1).
• Three radiologists independently reviewed compressed images to determine diagnostic yield.

Main Results:

• Clinical accuracy decreased gradually with increasing irreversible compression ratios.
• At a 7.4:1 compression ratio, sensitivity was 99% and specificity was 93%.
• At a 30.6:1 compression ratio, sensitivity dropped to 75% and specificity to 83%.
• Compression and decompression times decreased with higher compression levels.

Conclusions:

• Quadtree-based compression of abdominal CT images demonstrates potential for practical application in PC-based digital radiography.
• Moderate compression ratios offer a balance between data reduction and diagnostic performance.
• Further research may optimize compression parameters for clinical use.