Brain tumor is a life-threatening disease. Magnetic Resonance Imaging(MRI), is the investigative tool for locating tumor in brain. Deep Learning algorithms are being successfully implemented to support in medical imaging for the diagnosis of various diseases. Convolutional Neural Networks are DLalgorithms working remarkably on image classification and recognition. In this study, we classified images taken from brain MRI scans using the convolutional neural network (CNN) method. These images were divided into four categories (no tumor, glioma, pituitary, and meningioma). This research combines DL models with efficient ML classifiers to use both DL and ML techniques to their fullest. For experimentation, we applied Genetic Algorithm (GA) to develop a hybrid DL-GA-ML model. Comparisons are made between the proposed models and cutting-edge models. It was found that the proposed models achieved promising results in classifying brain tumor cases with excellent accuracy. The hybrid DL-GA-ML model outperforms compared approaches and yields an accuracy of 97.94%. Among Hybrid DL-ML models, the VGG16-Logistic Regression model attains an accuracy of 96.3%, the InceptionV3-Logitic Regression model 95.80%, whereas Resnet50-Logistic Regression exhibits 81.69% accuracy.

Deep Learning(DL), Brain Tumor, Machine Learning(ML), Genetic Algorithm(GA)