Estimation of Brain Tumor Volume Based on 3D Modelling for MRI Images

Abstract

Identifying the precise location, shape and size of the brain tumor has always been a task that newlinehas posed challenges due to brain s complex anatomy and its structure. Estimation of brain newlinetumor based on 3D modelling from MRI images will help radiologists and surgeons in this newlinechallenging task at their hand. This research undertaken in the domain of medical image newlineprocessing concentrates on developing segmentation algorithms with increased efficiency in newlineidentification of brain tumor region. It leads to better estimation of the brain tumor volume and newlineproviding a 3D model of the brain tumor to the radiologist and surgeons that can with faster newlinediagnosis, immediate treatment decisions for surgical planning leading to reduced morbidity newlineand mortality rate. newlineThe anatomical structures of the soft tissues in the human body are represented with high newlinecontrast values in the 2D MR image using Magnetic Resonance Imaging (MRI) system, a noninvasive newlineimaging technology. MR images of 50 patients suffering from brain tumor were newlinecollected based on inputs from radiologists from 3 different public and private hospitals. From newlinethis dataset 30 patient s datasets were shortlisted which satisfied both the inclusion criteria and newlineexclusion criteria. Consistency is ensured across the entire research work by utilizing the same newlineset of MR scans containing MR slice of 30 patients for results, comparison, analysis and newlinevalidation. To ensure uniformity in the results, comparison and further analysis, these 30 newlinepatients (subjects) and their MR image scans that are utilized are kept the same throughout the newlineresearch work. The radiologist were consulted while collecting the dataset to understand the newlineprocess of manual diagnostics that they practice which helped in planning the research work. newlineRadiologist s inputs were taken during the span of research work keeping in mind their need of newlineautomation in the diagnostics process. The results were also validated by the radiologist. newlineSegmentation of medical images conveys meaningful information to the radiologists.