Related Experiment Video
Updated: May 17, 2026
13:44
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
Published on: August 30, 2013
A texture analysis method for detection of clustered microcalcifications on digital mammograms
M C Barretto1, D A Kulkarni, G R Udupi
1Department of Information Technology, Padre Conceicao College of Engineering, Verna, 403712, Goa, India. christina.barretto@gmail.com
International Journal of Bioinformatics Research and Applications
|October 13, 2012
Summary
Early breast cancer detection is improved using a statistical texture analysis method to identify clustered microcalcifications on mammograms. This approach enhances diagnostic accuracy for early-stage breast cancer detection.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Biomedical Engineering
Background:
- Breast cancer remains a leading cause of mortality in women worldwide.
- Early detection through mammography, specifically identifying clustered microcalcifications, is crucial for improving patient outcomes.
- Texture analysis of mammographic images offers a promising avenue for automated detection of subtle abnormalities.
Purpose of the Study:
- To develop and evaluate a texture-based method for classifying regions of interest in mammograms.
- To differentiate between regions containing clustered microcalcifications (positive ROIs) and normal breast tissue (negative ROIs).
- To assess the efficacy of a backpropagation neural network classifier in breast cancer detection.
Main Methods:
- Utilized the Surrounding Region Dependence Method for statistical texture analysis based on second-order histograms.
- Extracted six distinct textural features from digitized mammograms.
- Employed a 3-layer backpropagation neural network for classifying regions of interest.
- Evaluated classification performance using Receiver Operating Characteristics (ROC) analysis.
Main Results:
- The implemented texture analysis method successfully extracted features capable of distinguishing between normal and abnormal breast tissue.
- The backpropagation neural network achieved a classification performance that was evaluated using ROC analysis.
- The study demonstrated the potential of texture analysis in enhancing the accuracy of microcalcification detection in mammograms.
Conclusions:
- Statistical texture analysis, specifically the Surrounding Region Dependence Method, is effective for identifying clustered microcalcifications in mammograms.
- The use of a backpropagation neural network classifier shows promise for automated breast cancer detection systems.
- This approach contributes to the advancement of early breast cancer detection techniques, potentially reducing mortality rates.
