Brain Tumor Detection Using ResNet Architectures

Any tumor occurs in the body due to uncontrolled and rapid growth of cells at a particular body part and then also affecting the healthy cells in the vicinity. On the off chance that not treated at an underlying stage, it might prove fatal. Despite many important efforts and promising results, accurate predictive testing and classification of these tumors is a daunting task. Important tests for the identification of this uncontrolled cell growth come from changes in the cancer's area, shape, and volume. The primary focus of this research is to convey a technique for early detection and on tumor growth recognition from MRIs so that the decision taking can be even more spontaneous in terms of detection and beginning of the treatment. This research aims at detecting tumors in a human brain depending on a dataset, which includes images of MRI scans of about 110 patients with lower-grade gliomas. A deep learning algorithmic approach, which is ResNet architecture, is implemented for detection of these tumors in the dataset. The model performed with significant accuracy levels following preprocessing and training with segmented masks of tumors.