Image compression in digital mammography: effects on computerized detection of subtle microcalcifications

Area of Science:

• Medical Imaging
• Digital Signal Processing
• Radiology

Background:

• Mammogram digitization at lower resolutions (e.g., 0.1 mm x 0.1 mm) significantly reduces radiologist detection accuracy for microcalcifications.
• Computerized detection accuracy also decreases with increasing pixel size, necessitating large matrix sizes for mammogram digitization to preserve image information.

Purpose of the Study:

• To evaluate the efficiency of two compression techniques, full frame Discrete Cosine Transform (DCT) with entropy coding and Laplacian Pyramid Hierarchical Coding (LPHC), for digital mammograms.
• To compare these techniques based on the trade-off between bit rate and detection accuracy of subtle microcalcifications by an automated algorithm.

Main Methods:

• Investigated the dependence of compression efficiency on compression parameters for both DCT and LPHC methods.
• Compared the techniques by analyzing the bit rate, detection accuracy of microcalcifications, mean-square errors, and visual quality of reconstructed images.

Main Results:

• Laplacian Pyramid Hierarchical Coding (LPHC) achieved a maximum compression ratio of 3.6:1 without significant degradation in microcalcification detectability.
• Full frame DCT with entropy coding offered a higher compression efficiency of 9.6:1 at comparable detection accuracy.
• Mean-square errors in reconstructed images did not correlate with the detection accuracy of microcalcifications.

Conclusions:

• The study highlights the importance of task-specific evaluation for assessing the quality of decompressed images in digital mammography.
• Full frame DCT with entropy coding demonstrates superior compression efficiency for digital mammograms while maintaining diagnostic accuracy.
• Further research is required to determine the optimal compression technique for digital mammograms based on specific clinical needs.