Mammography Image-Based Diagnosis of Breast Cancer Using Machine Learning: A Pilot Study.

Area of Science:

• Medical Imaging
• Machine Learning
• Oncology

Background:

• Breast tumors are common in women, requiring accurate classification.
• Mammogram analysis by radiologists is time-consuming and prone to errors.
• Machine learning offers potential for automated breast tumor classification.

Purpose of the Study:

• To develop and evaluate machine learning models for assisting radiologists in breast tumor classification from mammograms.
• To achieve high accuracy and efficiency in tumor classification.
• To identify optimal machine learning algorithms for this task.

Main Methods:

• Feature extraction from manually annotated regions of interest in mammograms.
• Training classification models using extracted features.
• Utilizing Support Vector Machine (SVM) and Naïve Bayes algorithms.
• Implementing feature selection and hyper-parameter optimization.

Main Results:

• The proposed system achieved 100% accuracy in classifying breast tumors.
• Optimized SVM and Naïve Bayes models demonstrated superior performance.
• The models provided classification within a reasonable time interval.

Conclusions:

• Machine learning-based techniques can significantly enhance breast tumor classification from mammograms.
• Optimized SVM and Naïve Bayes are highly effective for this application.
• The developed system assists radiologists, improving diagnostic efficiency and accuracy.