Computerized detection of clustered microcalcifications: a modular approach with non-linear filters

Area of Science:

• Medical Imaging
• Computer-Aided Diagnosis
• Digital Mammography

Background:

• The development of an ideal computerized mammogram processing system remains an ongoing challenge.
• Automated detection of microcalcifications is crucial for early breast cancer diagnosis.

Purpose of the Study:

• To propose a flexible, modular scheme for the automated detection and recognition of microcalcifications in mammograms.
• To introduce novel non-linear filters for enhanced microcalcification identification.

Main Methods:

• A modular processing scheme dividing the sequence into autonomous modules.
• Development of ROI selection modules utilizing specialized non-linear filters.
• Application of filters based on local statistical features and gradient component analysis.

Main Results:

• The proposed filters effectively select pixels with microcalcification-indicative statistical properties.
• Gradient-based filtering successfully eliminates noise and non-relevant sharp variations.
• Combined filter application enables precise identification of clustered microcalcifications.

Conclusions:

• The modular approach simplifies system maintenance and consistency in mammogram processing.
• This scheme facilitates the comparison of different processing techniques and parameters for microcalcification detection.
• The developed system offers a promising advancement for automated mammogram analysis.