Segmentation of Brain Tumor Objects in Magnetic Resonance Imaging (MRI) Image using Connected Component Label Algorithm
Prasetyo Mimboro, Andi Sunyoto, Rizqi Sukma Kharisma · 2021
Several techniques in computer vision can be used in object grouping, one of which is image segmentation. Segmentation is a field of computer vision that aims to group all pixels in an image into regions that have homogeneity in their characteristics. In this study, the segmentation technique will be carried out to partition objects in magnetic resonance image (MRI) images. Object segmentation uses a connected component labeling algorithm to classify regions or objects of brain tumors in MRI images. The dataset used in this study amounted to 80 images, of which 30 images contained meningioma brain tumor disease, 30 images with glioma brain tumor disease, and 20 images without indications of brain tumor disease. Based on tests conducted, the average accuracy obtained in detecting brain tumor objects is 96.66%.