Pre-trained CNN Models and Machine Learning Techniques for Brain Tumor Analysis

A brain tumor is a life-threatening condition that causes the brain's normal functions to be disrupted. For a definitive reading and an ef active treatment approach, brain tumors must be detected early. The physician's knowledge and experience determine a brain tumor's classification. However, an autonomous technique is necessary for rapid brain tumor analysis. The three extracting features algorithms VGG16, VGG19, and Inception v3 have been used in this study. Once the features were detected these are transferred to machine learning classifiers that distinguish between benign (non-cancerous) and malignant (cancerous) brain tumors. The results produce the highest accuracy of 97.7% using VGG16 and logistic regression classifier.