Intelligent deep learning approaches for automated mri based brain tumor detection

Abstract

Magnetic Resonance Imaging (MRI) modality is commonly used by radiologists in the medical field to diagnose abnormalities of the brain. The mortality rate due to Brain tumor is on the rise. The manual method of grading Brain tumors may lead to false negatives and false positives thus posing a challenge. The detection of abnormalities in the brain at an early stage is difficult too. Manual misprediction may reduce the chance of survival in people with brain abnormalities. Therefore, early and accurate detection of Brain tumors is essential. Hence, computer-aided diagnosis provides additional support to the radiologist in identifying brain tumor accurately at an early stage. Researchers have employed machine learning techniques to segment and classify brain tumors in MRI images. But due to the presence of intensity variations, noise and artifacts in the image, irregular nature of tumor characteristics and the computational complexity, these techniques are still not able to give accurate classification of images. Here, three different deep learning-based algorithms have been proposed to detect brain abnormalities from MRI modality. newline