The brain is the controlling unit for all functions in the human body. It controls functions like memory, vision, hearing, knowledge, personality, solving problems, etc. Brain tumors are caused by the uncontrolled growth of brain cells. In medical practices, accurate detection and identification of brain tumors is crucial. In literature, many techniques have been suggested by various researchers for detecting brain tumors. Magnetic resonance imaging (MRI) is used to generate high quality images of the brain. Brain tumors are diagnosed with the use of the MRI scans. In this paper we propose state of the art convolutional neural network models and use transfer learning to train image classification models and compare them. Techniques involved in image processing to detect a brain tumor consists of four stages: image augmentation, image segmentation, feature extraction, and classification.