MRI Brain Tumour Detection Using K-Nearest Neighbor Algorithm and Convolutional Neural Network Algorithm

Abstract

The primary goal of this research is to evaluate the Novel K-Nearest Neighbour (KNN) algorithm’s accuracy to that of the Convolutional Neural Network (CNN) method when detecting brain tumours in magnetic resonance images. Materials and Procedures to identify brain malignancies in MRI images, novel KNN and CNN algorithms were applied. A total of 40 samples of various MRI scans were gathered from a dataset that is accessible on Github. By gathering the dataset of 40 samples with 80% of pretest power, group 1 with 20 samples using KNN and group 2 with 20 samples using CNN were investigated. Results: According to the simulation results, the CNN algorithm and Novel KNN method both obtain accuracy levels of 84.10 percent and 98.5%, respectively. The significance level was 0.001 when calculated. When compared to the CNN algorithm, the KNN algorithm detects brain tumours with a substantially higher degree of accuracy.