BRAIN TUMOR ANALYSIS BY EXTRACTING FEATURES AND CLASSIFICATION THROUGH SUPPORT VECTOR MCHINES

Brain tumors are an increasing global epidemic, claiming millions of lives each year. Misdiagnosis can result in needless therapy and reduced life expectancy. Doctors have used computer-based diagnostic techniques such as DenseNet201 and the Gabor Filter to produce accurate diagnoses. In this work, SVM was used to classify independent features, and essential features were collected from an MRI image dataset using the DenseNet201 algorithm and Gabor filter. Deep convolutional layers outperform standard techniques in terms of extracting unique characteristics from target areas. An MRI dataset of 7023 brain tumor pictures from the Kaggle website was utilized to classify features using SVM. The hybrid approach of DenseNet201 and Gabor Filter produced the best overall results, with 98.02% precision, 98.01% accuracy, and 98.01% F1 score.