Bioinformatics, especially clinical imaging, has become a hot topic thanks to advances in artificial intelligence, particularly deep learning. It has been quite effective in assisting Computer-Aided Diagnosis (CAD) in achieving accurate outcomes. The detection of brain tumor Magnetic Resonance imaging, despite this, is still regarded as a major challenge. The brain tumor is a life-threatening threat that is only going to get worse. Detection at an early stage could minimize the risk of mortality. To detect brain tumors, researchers are now employing a variety of machine vision-based approaches. This work focuses on a combination technique for brain tumor detection that incorporates machine learning and deep learning. The research initiate with feature extraction employing a convolutional neural network (Alex Net) and continued with classification by introducing an ensemble classifier. Early detection and diagnosis of brain tumors employing a non-invasive, contactless machine vision system are being proposed. To develop a multiclass ensemble classification model, many statistical analyses were employed and conducted. Upon Comparison with other methods, the results show that the proposed method is 96 % efficient.

Index Terms- Glioma, Meningioma, Benign, Malignant, AlexNet, Ensemble classifier, Machine vision and Clinical image processing.