Computer-aided diagnosis of skin cancer based on soft computing techniques

Area of Science:

• Medical Imaging
• Artificial Intelligence in Healthcare
• Computational Biology

Background:

• Skin cancer poses a significant health risk, necessitating early detection for effective treatment.
• Current diagnostic methods can be time-consuming and may benefit from automated assistance.

Purpose of the Study:

• To develop and evaluate an automated computer-aided system for the early diagnosis of skin cancer.
• To enhance diagnostic accuracy through optimized image processing and machine learning techniques.

Main Methods:

• Image noise reduction using a median filter.
• Image segmentation via a convolutional neural network optimized with Satin Bowerbird Optimization (SBO).
• Feature extraction and selection using the SBO algorithm, followed by Support Vector Machine classification.

Main Results:

• The proposed system demonstrated high performance in classifying skin images.
• Comparative analysis showed the system's effectiveness against ten other methods.
• Key performance metrics included accuracy, sensitivity, specificity, and predictive values.

Conclusions:

• The developed computer-aided system shows promise for accurate and early skin cancer diagnosis.
• Optimized deep learning and machine learning approaches can significantly improve skin cancer detection rates.
• This automated method can support clinicians in identifying cancerous skin lesions.