The multi-level image segmentation in dermatology application using an enhance Secretary Bird Optimization Algorithm

Area of Science:

• Dermatological image analysis and artificial intelligence.
• Medical imaging and computational pathology.
• Optimization algorithms for biomedical applications.

Background:

• Dermatological diseases present global diagnostic and treatment challenges.
• High-resolution imaging and datasets like the Skin Condition Image Network (SCIN) are advancing dermatology.
• Accurate image segmentation is crucial for automated diagnosis, lesion identification, and measurement in dermatological images, but faces challenges due to variable skin textures, lighting, and lesion appearances.

Purpose of the Study:

• To enhance multilevel image segmentation optimization specifically for SCIN dermatological images.
• To improve segmentation accuracy, robustness to artifacts, and computational efficiency using novel optimization techniques.
• To investigate the clinical benefits of accurate segmentation in automated dermatological diagnostics and AI-driven healthcare solutions.

Main Methods:

• Development of an enhanced multilevel image segmentation optimization method tailored for SCIN dataset images.
• Integration of Opposition-Based Learning (OBL) and Orthogonal Learning (OL) into the Secretary Bird Optimization Algorithm (SBOA), creating the mSBOA.
• Application of the mSBOA to address challenges like overlapping textures, variable illumination, and image artifacts in dermatological images.

Main Results:

• The enhanced mSBOA demonstrates increased segmentation accuracy for dermatological images.
• Improved robustness against common image artifacts and variable lighting conditions was observed.
• The optimized algorithm achieved higher computational efficiency without compromising diagnostic accuracy.

Conclusions:

• The proposed enhanced optimization method significantly improves multilevel feature segmentation in SCIN dermatological images.
• Accurate segmentation facilitates personalized treatment, enhances patient outcomes, and reduces diagnostic errors.
• This advancement supports the wider adoption of AI in dermatology, particularly in underserved areas, by improving automated diagnostic capabilities.