Skin Cancer Detection Using Kernel Fuzzy C-Means and Improved Neural Network Optimization Algorithm

Area of Science:

• Medical Imaging
• Computational Pathology
• Artificial Intelligence in Oncology

Background:

• Early diagnosis of malignant skin cancer is crucial for effective treatment.
• Computer-aided diagnosis (CAD) systems offer potential for improving diagnostic accuracy and efficiency.

Purpose of the Study:

• To propose an optimized pipeline methodology for computer-aided diagnosis of skin cancers from images.
• To develop and evaluate a novel metaheuristic algorithm for optimizing feature selection and classification stages.

Main Methods:

• The pipeline involves image preprocessing (noise reduction, contrast enhancement), region of interest (ROI) segmentation using kernel fuzzy C-means, feature extraction, and feature selection.
• A developed neural network optimization algorithm (a metaheuristic) was employed to optimize feature selection and Support Vector Machine (SVM) classification.
• The system was validated through comparison with five state-of-the-art methods.

Main Results:

• The proposed pipeline demonstrated superior performance compared to existing state-of-the-art methods.
• The integration of the novel metaheuristic algorithm significantly enhanced the optimization of feature selection and SVM classifier accuracy.

Conclusions:

• The developed computer-aided diagnosis pipeline offers a highly effective approach for the early detection of skin cancer.
• The novel optimization algorithm shows promise for improving the performance of CAD systems in medical image analysis.