Brain Tumor Detection and Classification Using an Optimized Convolutional Neural Network.

Area of Science:

• Medical Imaging and Artificial Intelligence
• Computational Neuroscience
• Oncology

Background:

• Brain tumors are a significant global health challenge, leading to high mortality rates.
• Accurate and timely diagnosis is crucial for effective brain tumor treatment and patient survival.
• Existing diagnostic methods can be improved with advanced computational approaches.

Purpose of the Study:

• To develop and optimize a hyperparametric convolutional neural network (CNN) model for enhanced brain tumor identification.
• To systematically fine-tune CNN hyperparameters to improve feature extraction and reduce model complexity.
• To provide a more accurate and effective tool for medical practitioners in diagnosing brain tumors.

Main Methods:

• Utilized a hyperparametric convolutional neural network (CNN) architecture.
• Fine-tuned critical hyperparameters including batch size, layer count, learning rate, activation functions, pooling, padding, and filter size.
• Trained and validated the model on three diverse brain MRI datasets from Kaggle.

Main Results:

• Achieved outstanding performance with an average score of 97% for accuracy, precision, recall, and F1-score across datasets.
• Demonstrated superior performance compared to state-of-the-art approaches through methodical comparisons.
• Enhanced model generalization capacity through hyperparameter optimization.

Conclusions:

• The hyperparameter-tuned CNN model offers a significant advancement in accurate and trustworthy brain tumor diagnosis.
• The optimized model provides medical practitioners with a more effective tool for critical diagnostic judgments.
• This research has practical implications for improving patient outcomes in brain cancer cases.