Detection of Lung Cancer Cells Using Inception V3 Model

Abstract

Lung Cancer is a dangerous disease which is caused by consumption of tobacco and other pollutants it is very hard to predict. But in this era of technology, prediction of lung cancer is possible with the help of machine learning and deep learning algorithms. According to the surveys the survival rate of patients that are affected by cancer is very low, it is because of lack of prediction of cancerous cells in earlier phase. In this Project we use CT(Computed Tomography) scan images of malignant, benign and normal cases are taken in the dataset to classify them as cancerous and non-cancerous images with the help of CNN family Algorithm, i.e, Inception V3 Algorithm. The CT scan images are preprocessed with the help of image preprocessing techniques such as image warping and image cropping. After the preprocessing of the CT images the images are processed with Inception V3 Algorithm, the image dataset is trained under Inception V3 Algorithm of deep learning technique and under the range of 10 epochs the best accuracy the algorithm has provided is 97%. Many Machine Learning Algorithms are used to classify the images such as SVM, CNN, etc. But under deep learning technique the inception V3 algorithm has provided the best accuracy so far. Keywords— CT Scan Images, CNN Algorithm, Cancer Nodules, Malignant cases, Benign Cases, Normal Cases, Inception V3 Algorithm, SVM Algorithm, Image Warping and Image Cropping.