Particle Swarm Optimization and Two-Way Fixed-Effects Analysis of Variance for Efficient Brain Tumor Segmentation.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Biomedical Engineering

Background:

• Accurate brain tumor segmentation is crucial for diagnosis and treatment planning in medical imaging.
• Existing segmentation methods face challenges in precisely delineating abnormal masses in MRI scans.
• Developing robust and efficient automated segmentation techniques is an ongoing research priority.

Purpose of the Study:

• To propose a novel, optimized segmentation method for detecting brain lesions in MRI using particle swarm optimization (PSO).
• To enhance the accuracy and robustness of brain tumor detection and classification.
• To evaluate the proposed method against state-of-the-art techniques using established performance metrics.

Main Methods:

• A three-step segmentation process involving skull bone removal, PSO-based lesion block detection, and K-means classification.
• Utilizing a two-way fixed-effects analysis of variance (ANOVA) as the fitness function for PSO.
• Validation on private MRI and BraTS 2015 datasets, comparing ANOVA fitness to sum-of-absolute-differences (SAD).

Main Results:

• The proposed ANOVA-based fitness function demonstrated superior efficiency and robustness compared to SAD.
• The optimized segmentation algorithm significantly outperformed several state-of-the-art techniques.
• Performance was validated using Dice coefficient, Jaccard distance, correlation coefficient, and RMSE metrics.

Conclusions:

• The developed optimized segmentation algorithm shows significant superiority for brain tumor detection in MRI.
• The integration of PSO and ANOVA offers a promising approach for enhancing medical image analysis.
• This method provides a robust and accurate tool for refining the understanding of brain lesions.