A grade-based search adaptive random slime mould optimizer for lupus nephritis image segmentation

Area of Science:

• Medical Image Processing
• Computational Intelligence
• Biomedical Engineering

Background:

• Medical image segmentation is vital for data analysis but faces challenges with traditional multi-threshold techniques.
• Existing methods are computationally intensive and yield suboptimal results, limiting their practical application.

Purpose of the Study:

• To develop an advanced multi-threshold image segmentation algorithm for improved medical image analysis.
• To enhance the slime mould algorithm (SMA) with multiple strategies for better performance.

Main Methods:

• Developed a novel multi-strategy-driven slime mould algorithm (RWGSMA).
• Incorporated random spare, double adaptive weight, and grade-based search strategies to optimize SMA.
• Evaluated RWGSMA using IEEE CEC2017 test suites and typical medical images.
• Applied RWGSMA with 2D Kapur's entropy for lupus nephritis histopathological image segmentation.

Main Results:

• RWGSMA demonstrated accelerated convergence and avoidance of local optima.
• The algorithm showed significant improvements in segmentation performance on test suites and medical images.
• RWGSMA outperformed several competing algorithms in segmenting lupus nephritis images.

Conclusions:

• The proposed RWGSMA offers a robust and efficient solution for multi-threshold medical image segmentation.
• RWGSMA shows significant promise for applications in histopathological image analysis, particularly for diseases like lupus nephritis.