Breast Cancer Classification Using Feature Selection via Improved Simulated Annealing and SVM Classifier.

Area of Science:

• Medical Imaging
• Artificial Intelligence in Healthcare
• Oncology

Background:

• Breast cancer is a leading cause of mortality in women, necessitating early detection for improved outcomes.
• Computer-aided diagnostic (CAD) systems are vital for enhancing mammogram analysis accuracy and expediting clinical decisions.

Purpose of the Study:

• To develop an automated CAD system for detecting cancerous tumors in mammograms.
• To improve the accuracy and efficiency of breast cancer diagnosis through advanced feature selection and classification techniques.

Main Methods:

• An automated CAD system comprising preprocessing, feature extraction, feature selection (Improved Simulated Annealing - ISA), and classification (Support Vector Machine - SVM).
• Preprocessing involved ROI extraction, noise suppression, and contrast enhancement.
• ISA algorithm adaptively selected informative features using a composite fitness function to reduce dimensionality while maintaining accuracy.

Main Results:

• The system achieved high performance on CBIS-DDSM and MIAS datasets, with accuracies of 99.67% and 98%, respectively.
• Sensitivities reached 99.33% and 98%, and F1-scores were 99.66% and 97.9%, respectively.
• The ISA algorithm demonstrated superior performance compared to traditional Simulated Annealing and full-feature methods.

Conclusions:

• The ISA algorithm is effective in selecting relevant features for breast cancer detection.
• The proposed CAD system significantly enhances diagnostic performance in mammogram analysis.