Brain Tumor Recognition Leveraging Machine Learning and CNN

A mass or lump that can be formed by an abnormal growth of cells in the body is known as a tumor. Tumors are of two types: Non-cancerous or cancerous. Cells in the human brain essentially grow in a distorted manner, causing a brain tumor. A tumor in the brain is the harmful disease and can be analyzed by magnetic resonance imaging (MRI) images using automated techniques. Many machine learning (ML) methods for identifying and analyzing brain tumors have been proposed by several researchers so far. A comparative study of different methods was carried out. After an intensive analysis process, some issues were discovered in image segmentation, recovery, and enhancement. To sort out these problems, a new method is proposed by using convolutional neural network (CNN). CNN has the ability to identify complex patterns and relationships in images. CNN algorithms can accurately analyze large amounts of data such as MRI images to identify the presence of a tumor, its size, location, and type. To reduce the errors in images of the dataset, CNN classification is used. CNNs are superior to other traditional image analysis algorithms such as threshold or edge detection. To implement the proposed model, Python, TensorFlow, and Matplotlib environments are used. CNN can predict tumors in MRI images with high accuracy. The proposed method using CNN outperforms deep learning (DL), genetic algorithm (GA), support vector machine (SVM), and artificial neural network (ANN) and gives the validation accuracy up to 95.7%.