Prediction of Automatic Lung Tumor Segmentation using X-Ray Images

Abstract

Pulmonary cancer also known as lung carcinoma is the leading cause for cancer-related death in the world. Early stage cancer detection using x-ray technics could save hundreds of thousands of lives every year. However analysing hundreds of thousands of these scans are an enormous burden for radiologists and too often they suffer from observer fatigue which can reduce their performance. Therefore, a need to read, detect and provide an evaluation of x-rays efficiently exists. Detection of exact tumor of lung can be done by lung tumor segmentation and accurate longitudinal tracking of tumor volume changes from x-ray images. Here we are proposing a method of detecting lung cancer in a x-ray using a 2D-UNet model . We cropped 2D cancer masks on its reference image using the center of the lung cancer given in the dataset and trained a model with different machine learning techniques and hyperparameters. In summary we have a developed an approach for volumetrically segmenting lung tumors which enables accurate, automated identification of the serial measurement of tumor volumes in the lung.