A novel minimum generalized cross entropy-based multilevel segmentation technique for the brain MRI/dermoscopic...

Area of Science:

• Medical Image Analysis
• Computational Intelligence
• Biomedical Engineering

Background:

• Accurate disease source identification in medical imaging is crucial for early diagnosis and treatment.
• Challenges in medical image analysis include aberrant tissue changes that can lead to life-threatening conditions like cancer.
• Efficient automatic image segmentation is essential for reliable medical image interpretation.

Purpose of the Study:

• To enhance the efficiency and reliability of medical image segmentation processes.
• To develop an optimal objective function and an effective optimization algorithm for segmentation.
• To address the limitations of existing segmentation techniques in medical image analysis.

Main Methods:

• Introduced a novel objective function: minimum generalized cross entropy (MGCE).
• Developed a new optimization algorithm: opposition African vulture optimization algorithm (OAVOA).
• Incorporated opposition-based learning into the OAVOA to improve exploration capabilities.

Main Results:

• The OAVOA demonstrated superior performance compared to state-of-the-art optimizers on benchmark functions.
• The OAVOA-MGCE multilevel thresholding approach achieved superior results on Brain MRI and dermoscopic images.
• Comparative analysis confirmed the effectiveness of the proposed method over other entropy-based thresholding techniques.

Conclusions:

• The proposed OAVOA-MGCE method offers a significant advancement in medical image segmentation.
• This approach provides a more accurate and reliable tool for analyzing medical images.
• The developed optimization algorithm and objective function hold promise for future medical image analysis applications.