Brain tumor recognition by an optimized deep network utilizing ammended grasshopper optimization.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Neuroscience

Background:

• Brain tumors, abnormal cell masses, require timely detection for effective treatment.
• Accurate diagnosis is challenged by tumor heterogeneity and location.
• Current diagnostic methods can be time-consuming and require expert interpretation.

Purpose of the Study:

• To develop and evaluate a novel deep learning approach for accurate and rapid brain tumor diagnosis using MRI.
• To enhance the efficiency and accuracy of brain tumor classification through a hybrid AI model.
• To assess the proposed method's performance against existing state-of-the-art techniques.

Main Methods:

• Feature extraction from brain MRI images using AlexNet.
• Complexity reduction of AlexNet by integrating an Extreme Learning Machine (ELM) as a classification layer.
• Parameter optimization of the ELM network using an Amended Grasshopper Optimization Algorithm (AGOA).

Main Results:

• The proposed method achieved high performance metrics: 0.96 accuracy, 0.94 precision, 0.96 specificity, 0.96 F1-score, 0.94 sensitivity, and 0.90 MCC.
• The model demonstrated robustness and stability across varying noise levels and image resolutions.
• The approach outperformed several state-of-the-art techniques in brain tumor classification.

Conclusions:

• The developed deep learning approach offers a rapid, accurate, and dependable method for brain tumor diagnosis.
• The hybrid model integrating deep learning and meta-heuristic optimization shows significant potential for medical image analysis.
• This method could be valuable for improving clinical decision-making in neuro-oncology.