Brain Abnormality Classification by Combination of Machine Learning and Deep Learning
Keywords:
Convolutional Neural Network (CNN), Deep Learning (DL), Glioma, Grayscale, K-Nearest Neighbours (KNN), Machine Learning (ML), Magnetic Resonance Imaging (MRI), Meningioma, Normalization, Pituitary, Stroke, TumorAbstract
Brain abnormalities, such as tumours and strokes, by taking its average MRI data and their early detection are paramount for timely intervention and effective treatment. This research describes a novel method for identifying a variety of brain abnormalities, such as tumours and strokes, in medical imaging data obtained from Magnetic Resonance Imaging (MRI) scans by combining the strengths of Deep Learning (DL) and Machine Learning (ML). The proposed methodology integrates DL and ML to provide a comprehensive and accurate diagnostic tool. Convolutional Neural Networks (CNNs) are employed for feature extraction, enabling the model to automatically identify intricate patterns and abnormalities within the brain images. The extracted features are subsequently fed into a carefully designed algorithm of ML classifiers, k-Nearest Neighbors (KNN). This ensemble approach capitalizes on the strengths of ML to make precise diagnostic decisions based on the rich feature representations derived from the DL component. Our model has achieved more remarkable accuracy and is very efficient.
