An optimal fast fractal method for breast masses diagnosis using machine learning.

Area of Science:

• Medical Imaging
• Computational Biology
• Oncology

Background:

• Fractal analysis is useful for medical image analysis but computationally intensive.
• Classifying breast lesions in mammography requires efficient and accurate methods.
• Existing fractal methods face challenges with computational load and processing time.

Purpose of the Study:

• To introduce a computationally efficient fractal method for breast lesion classification in mammography.
• To enhance classification accuracy by optimizing fractal information extraction.
• To reduce the computational burden associated with fractal analysis in medical imaging.

Main Methods:

• Developed a fast fractal method for breast lesion classification.
• Defined an objective function to identify the optimal fractal scale for classification.
• Extracted and utilized fractal information exclusively from the best identified scale.
• Validated the method using Support Vector Machine (SVM), Genetic Algorithm (GA), and Deep Learning (DL) classifiers.

Main Results:

• The proposed method significantly improves computation speed and reduces processing load.
• Classification accuracy is enhanced by focusing on optimal fractal information.
• Validation across SVM, GA, and DL confirmed the method's effectiveness.
• Comparative analysis showed superior classification performance compared to existing studies.

Conclusions:

• Optimizing fractal information extraction by selecting the best scale is crucial for efficient and accurate breast lesion classification.
• The proposed fast fractal method offers a promising approach for improving mammographic analysis.
• This technique has the potential to enhance diagnostic capabilities in breast cancer screening.