Kapur's entropy for multilevel thresholding image segmentation based on moth-flame optimization

Area of Science:

• Computer Vision
• Artificial Intelligence
• Image Processing

Background:

• Multilevel thresholding is crucial for image segmentation but suffers from increased computational complexity and reduced accuracy with higher threshold levels.
• Existing methods struggle to balance exploration and exploitation, leading to suboptimal segmentation results.

Purpose of the Study:

• To introduce a novel Moth-Flame Optimization (MFO) algorithm for multilevel thresholding image segmentation.
• To address the limitations of traditional methods by enhancing computational efficiency and segmentation accuracy.

Main Methods:

• The study proposes a Moth-Flame Optimization (MFO) algorithm integrated with Kapur's entropy for multilevel thresholding.
• MFO's exploration and exploitation capabilities are adjusted to optimize fitness values for segmentation.
• Performance is validated by comparing MFO against other optimization algorithms using critical evaluation indicators.

Main Results:

• Moth-Flame Optimization (MFO) demonstrates a faster convergence speed compared to other algorithms.
• The proposed MFO method achieves higher calculation accuracy and a superior segmentation effect.
• Experimental results indicate that MFO offers better stability in multilevel thresholding image segmentation.

Conclusions:

• Moth-Flame Optimization (MFO) provides an effective and efficient solution for multilevel thresholding image segmentation.
• The MFO algorithm overcomes the computational complexity and accuracy limitations of traditional methods.
• MFO is a promising technique for achieving high-quality image segmentation with improved performance metrics.