Related Experiment Video
Updated: Sep 4, 2025
13:44
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
Published on: August 30, 2013
43.0K
Classification of Mammographic ROI for Microcalcification Detection Using Multifractal Approach.
Nadia Kermouni Serradj1, Mahammed Messadi2, Sihem Lazzouni2
1Biomedical Engineering Laboratory, Faculty of Technology, Abou Bekr Belkaid University, 13000, Tlemcen, Algeria. nadia.kermouni@gmail.com.
Journal of Digital Imaging
|July 19, 2022
Summary
This study introduces a multifractal approach for detecting microcalcifications (MCs) in mammograms, aiming to identify precancerous cells. The developed system achieved high accuracy, demonstrating its potential for improved breast cancer screening.
Area of Science:
- Medical Imaging
- Computational Biology
- Biophysics
Background:
- Microcalcifications (MCs) are critical indicators of precancerous cells in mammography.
- Accurate detection of MCs is essential for early breast cancer diagnosis and improved patient outcomes.
- Developing automated systems for MC detection presents a significant research challenge.
Purpose of the Study:
- To propose and evaluate a novel computer-aided detection (CAD) system for microcalcifications (MCs) in mammographic images.
- To utilize a multifractal approach for classifying regions of interest (ROIs) as normal or abnormal (containing MCs).
Main Methods:
- A four-step methodology involving mammogram pre-processing (breast selection, haze removal, contrast enhancement).
- Extraction of ROIs, calculation of multifractal spectra, and extraction of multifractal and GLCM features.
- Classification of ROIs using K-Nearest Neighbors (KNN), Decision Trees (DT), and Support Vector Machines (SVM) classifiers.
Main Results:
- The system was evaluated on the INbreast database, analyzing 2688 ROIs.
- The Support Vector Machines (SVM) classifier achieved the highest performance.
- SVM yielded a sensitivity of 98.66%, specificity of 97.77%, and precision of 98.20%.
Conclusions:
- The proposed multifractal-based system demonstrates high efficacy in detecting microcalcifications (MCs).
- The SVM classifier proved most effective, offering excellent diagnostic performance for breast cancer screening.
- These results are competitive with existing literature, highlighting the system's potential clinical utility.
