Related Experiment Videos
Segmentation of suspicious clustered microcalcifications in mammograms
M A Gavrielides1, J Y Lo, R Vargas-Voracek
1Department of Biomedical Engineering, Duke University, Durham, North Carolina 27708, USA. marios@duke.edu
Medical Physics
|February 5, 2000
Summary
A new computer-aided diagnosis (CAD) scheme accurately segments suspicious microcalcification clusters in mammograms. This automated tool shows promise for improving early breast cancer detection efficiency and accuracy.
Area of Science:
- Medical Imaging
- Computer-Aided Diagnosis
- Breast Cancer Detection
Background:
- Early detection of breast cancer is crucial for effective treatment.
- Microcalcification clusters are important indicators of early breast cancer.
- Automated analysis of mammograms can aid radiologists in identifying suspicious regions.
Purpose of the Study:
- To develop and evaluate a multistage computer-aided diagnosis (CAD) scheme.
- To automate the segmentation of suspicious microcalcification clusters in digital mammograms.
- To improve the accuracy and efficiency of breast cancer diagnosis.
Main Methods:
- A three-step CAD scheme involving breast region segmentation, enhancement, and microcalcification detection.
- Utilized unsharp masking for high-frequency enhancement and local histogram analysis with fuzzy logic for segmentation.
- Employed fuzzy logic rules based on cluster features to differentiate benign from suspicious clusters.
Main Results:
- Achieved a true positive rate of 93.2% for microcalcification cluster segmentation.
- Reported an average of 0.73 false positive clusters per image.
- Demonstrated comparable sensitivity with an improved false positive rate compared to existing methods.
Conclusions:
- The developed CAD scheme effectively automates the segmentation of suspicious microcalcification clusters.
- The system shows encouraging performance for accurate and efficient breast cancer diagnosis.
- This automated tool has the potential to assist radiologists in mammogram interpretation.