Related Experiment Video
Updated: Mar 30, 2026
13:44
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
Published on: August 30, 2013
43.9K
Ensemble Supervised Classification Method Using the Regions of Interest and Grey Level Co-Occurrence Matrices
Hossein Yousefi Banaem1, Alireza Mehri Dehnavi2, Makhtum Shahnazi3
1Department of Biomedical Engineering, Faculty of Advanced Medical Technology, Isfahan University of Medical Sciences, Isfahan, Iran.
Summary
This study introduces an ensemble classification method for detecting breast cancer from mammograms. The new tool achieved high accuracy, aiding radiologists in identifying abnormal cases and improving diagnostic precision.
Area of Science:
- Medical Imaging
- Computational Pathology
- Machine Learning in Healthcare
Background:
- Breast cancer is a prevalent cancer in women, posing diagnostic challenges.
- Accurate classification of mammograms into malignant or benign is crucial for patient outcomes.
Purpose of the Study:
- To classify mammogram data as normal or abnormal using an ensemble classification method.
- To develop a computer-aided diagnostic tool for breast cancer detection.
Main Methods:
- Texture features were extracted from mammograms using Gray-Level Co-occurrence Matrices (GLCM).
- Feature selection was performed using the maximum difference method to identify distinguishing features.
- An ensemble supervised algorithm, combining three classifiers with a voting policy, was employed for classification.
Main Results:
- The proposed ensemble method achieved high diagnostic performance.
- Sensitivity was recorded at 96.66% and specificity at 97.50% using a perfect test method.
Conclusions:
- The developed computer-aided diagnostic tool demonstrates reliability in breast cancer detection.
- The method assists radiologists by improving diagnostic accuracy for abnormal mammogram data.
