Classification Technique for Brain Tissues Diagnoses: Segmenting Healthy, Cancer Affected and Edema Brain Tissues

Abstract

Brain tumors, are the most common and aggressive disease, leading to a very short life expectancy in their highest grade. Thus, treatment planning is crucial to improving patients’ quality of life. Generally, various image techniques such as Computed Tomography (CT), Magnetic Resonance Imaging (MRI), and ultrasound image are used to evaluate the tumor in a brain, lung, liver, breast, prostate, etc. Primarily, in this work, MRI images are used to diagnose a tumor in the brain. However, MRI scans’ huge amount of data thwarts manual classification of tumor vs non-tumor at a particular time. However, with some limitations, accurate quantitative measurements are provided for a limited number of images. Hence, trusted and automatic classification schemes are essential to prevent the human death rate. The automatic brain tumor classification is a very challenging task in large spatial and structural variability of the surrounding brain tumor region. This work proposes automatic brain tumor detection to segment the Region Proposal Network (RPN) by the Faster R-CNN algorithm. Here, the concept of transfer learning is used during training. The proposed system helps predict the correct type of tumor with better accuracy, about 99%, and classifies using Convolutional Neural Networks (CNN). The deeper architecture design is performed by using small kernels. Experimental results show that the CNN archives rate of 98% accuracy with low complexity compared with all other states of arts.