A novel computer aided breast mass detection scheme based on morphological enhancement and SLIC superpixel...

Area of Science:

• Medical Imaging
• Computer-Aided Diagnosis
• Biomedical Engineering

Background:

• Mammography is crucial for early breast cancer detection.
• Computer-aided detection (CAD) systems aim to improve mammogram interpretation accuracy.
• Reducing false positives (FP) in CAD systems is essential for efficient screening.

Purpose of the Study:

• To develop a novel computer-aided detection (CAD) scheme for mass detection on digitized mammograms.
• To achieve high sensitivity in mass detection while maintaining a low false positive (FP) rate.
• To leverage morphological enhancement and simple linear iterative clustering (SLIC) for improved performance.

Main Methods:

• A multi-stage CAD approach incorporating morphological enhancement for pattern highlighting and background clutter removal.
• Segmentation of mass candidates using the simple linear iterative clustering (SLIC) method.
• A pipeline including rule-based prescreening, distance regularized level set evolution for contour refinement, and ensemble support vector machines for FP reduction.

Main Results:

• The CAD system demonstrated high mass-based sensitivity, reaching 98.55% at 0.84 FP/image on the test dataset.
• Excellent performance was observed on a separate non-mass dataset, with FP rates as low as 0.30 FP/image.
• The system effectively reduced false positives while maintaining high detection rates for breast masses.

Conclusions:

• The developed CAD system shows significant promise for enhancing breast cancer screening efficacy.
• The integration of morphological enhancement and SLIC segmentation contributes to improved CAD system performance.
• The proposed CAD scheme offers a valuable tool for reducing false positives and increasing sensitivity in mammographic analysis.