Implementation of a model for detection and classification of brain tumours in magnetic resonance imaging using convolutional neural networks

Abstract

Accurate detection and classification of brain tumours in magnetic resonance imaging (MRI) are crucial for diagnosis and treatment planning. This research paper presents the implementation of a comprehensive model for the detection and classification of brain tumours using convolutional neural networks (CNNs) based on T1-weighted MRI scans. The project encompasses the development of a data preprocessing pipeline, including data normalisation, train/validation/test set splitting, and organisation into a suitable directory structure. The pipeline ensures the creation of a balanced and representative dataset for training and evaluating the CNN-based tumour classification model. The tumour detection and classification algorithm utilize CNNs to analyse preprocessed T1-weighted MRI data. The 3D CNN model leverages the spatial information encoded in the MRI volumes to accurately identify and classify brain tumours. TensorFlow, a popular deep learning library, is employed for developing and training the 3D CNN model. The model's performance is evaluated using appropriate metrics such as accuracy, precision, and area under the ROC curve (AUC). The results demonstrate the effectiveness of the proposed model in detecting and classifying brain tumours in T1-weighted MRI scans, with high accuracy and discriminatory power. Overall, the implementation of this model for brain tumour detection and classification in T1-weighted MRI scans provides a valuable tool for medical professionals and researchers. The model's accuracy and efficiency contribute to improved diagnosis, treatment planning, and monitoring of brain tumours, ultimately enhancing patient care and outcomes.