Volume 13 | Issue 4
Volume 13 | Issue 4
Volume 13 | Issue 4
Volume 13 | Issue 4
Volume 13 | Issue 4
Brain tumour segmentation is an especially difficult but important task in medical imaging. This is due to the possibility of incorrect diagnosis and prognosis resulting from manual classification. When dealing with large amounts of data, it also takes a lot of labour. Photographically, distinguishing between a brain tumour and normal tissue may be challenging due to their comparable looks and vast range of characteristics. In this work, brain tumours were extracted from two-dimensional MRI images using fuzzy C-Means. Classical classifiers and CNN were then used. The dataset utilised in the research included tumours of different sizes, forms, and intensities. SVM, KNN, MLP, LR, Naive Bayes, Random Forest, and other well-known machine learning algorithms were employed in the standard classifier of the scikit-learn module. Next, convolutional neural networks (CNNs), which are more adept at forecasting results than ordinary neural networks created with Keras or Tensorflow, were studied. CNN's accuracy rate, according to our study, was 97.87%. This study's main objective is to distinguish between normal and aberrant pixels using statistical and textural data