An Improved AI Inspired Brain Tumor Detection

This research paper aims to propose a sophisticated AI technique namely Deep Neural Network (DNN) and Support Vector Machine (SVM) to extract and detect brain tumors. The preternatural growth in brain leading towards brain tumor requires proper assessment and timely recognition for further diagnosis. Magnetic Resonance Imaging (MRI) is effectively significant to achieve accuracy and efficiency in the overall treatment. The dimensions of MRI Images are resized as required for AlexNet input images that are further utilized in the processing stage and it is evident from MRI scans available in the manuscript. The approach used for the extraction of features is based on the application of image enhancement using Contrast Limited Adaptive Histogram Equalization (CLAHE) technique . The process is started by the initial processing of MRI images and application of grayscale in the preprocessing stage. The data input is eventually differentiated by adding extra commotion to the picture using median and morphology filter. The features after being converted into grayscale are extracted from this high-quality refined image with the aid of the AlexNet model and these are then moved to SVM and k-Nearest Neighbor (KNN) for identification of tumor. Finally, an exact position of the tumor is obtained and therefore it becomes easier to diagnose the type and remove the tumor.