Brain-GCN-Net: Graph-Convolutional Neural Network for brain tumor identification.

Area of Science:

• Artificial Intelligence
• Medical Image Analysis
• Deep Learning

Background:

• AI and medical imaging are revolutionizing brain tumor detection and diagnosis.
• Accurate detection and classification are vital for early diagnosis and treatment.
• Traditional Convolutional Neural Network (CNN) models struggle with complex medical image features.

Purpose of the Study:

• To propose a fused Deep Learning (DL) model integrating Graph Neural Networks (GNN) and CNN for improved brain tumor detection.
• To enhance the comprehensive representation of brain tumor images and boost classification performance.

Main Methods:

• A fused DL model combining GNNs for relational dependencies and CNNs for spatial features was developed.
• The model was evaluated on a public dataset comprising 10,847 MRI images.
• Performance was compared against existing pre-trained models and traditional CNN architectures.

Main Results:

• The fused DL model achieved 93.68% accuracy in brain tumor classification.
• The proposed model demonstrated superior performance compared to existing pre-trained and traditional CNN models.
• This indicates the effectiveness of integrating GNNs and CNNs for this task.

Conclusions:

• The developed fused DL model shows significant potential for brain tumor classification.
• Further investigation in clinical trials is recommended to assess its utility in improving clinical decision-making.