An efficient framework for MRI brain Tumor classification using saliency driven non linear diffusion filtering and deep convolutional neural Networks

Abstract

An innovative imaging technology facilitates researchers to characterize, measure, and visualize the molecular and biological phenomena with unattainable accuracy. Imaging technologies allow breakthrough discoveries in various health-care applications and act as a centralized tool for biomedical researchers. In the last few decades, technological innovation in biomedical imaging has been growing with significant advancements. Thus, these advancements accelerate the call for the adoption of various new methodologies and care for the betterment of patientand#8223;s disease prediction in an earlier stage. newlineThis dissertation concentrates on MRI brain tumor classification using image processing. The brain tumor is a mass of tissues that are structured by anamalous cells and it is essential to classify brain tumors for further treatment. The manual investigation is a regular technique for MRI brain tumor detection and classification. However, it leads to complexities and lacks to give better solutions during the decision making process. Therefore, efficient tumor identification, extraction and classification are some challenging tasks for radiologists and physicians. This research attempts to give a better solution for the timely prediction of brain tumors using advanced learning approaches. newlineIn this research work, an automated MRI brain tumor prediction model is implemented with suitable image processing techniques and deep learning approaches to improve the prediction accuracy and to ease the process of diagnosing by physicians. newline