BTC-fCNN: Fast Convolution Neural Network for Multi-class Brain Tumor Classification.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Neuro-oncology

Background:

• Accurate brain tumor prognosis is vital for treatment planning.
• Manual classification of brain tumors in MRI images is complex and time-consuming.
• Automated computer-aided diagnosis (CAD) systems are needed for efficient and accurate tumor classification.

Purpose of the Study:

• To develop a fast and efficient deep learning-based system for classifying three types of brain tumors (meningioma, glioma, pituitary) from MRI images.
• To improve upon existing classification methods that suffer from performance limitations and high computational costs.

Main Methods:

• A novel deep learning model, BTC-fCNN, was proposed, featuring a 13-layer architecture with efficient components like convolution and pooling layers.
• The model was trained and validated on the Figshare dataset of MRI images.
• Transfer learning and five-fold cross-validation techniques were employed to enhance model performance.

Main Results:

• The BTC-fCNN model achieved high average accuracy, reaching 98.63% with transfer learning and 98.86% with cross-validation.
• The system demonstrated superior performance compared to state-of-the-art methods and other well-known convolutional neural networks (CNNs).
• Comprehensive evaluation metrics including precision, recall, F-score, and specificity confirmed the model's effectiveness.

Conclusions:

• The proposed BTC-fCNN system offers a fast, efficient, and highly accurate solution for automated brain tumor classification from MRI data.
• This deep learning approach holds significant potential for improving computer-aided diagnosis systems in neuro-oncology.
• The model's performance surpasses current state-of-the-art techniques, paving the way for more reliable tumor prognosis.