Pulmonary Cancer Prediction Using Machine Learning

Abstract

This project presents a Lung Cancer Detection 
System using Deep Learning and Graphical User 
Interface (GUI) implementation. The system is 
designed to import, preprocess, train, and evaluate 
lung cancer data obtained from DICOM (Digital 
Imaging and Communications in Medicine) images. 
It utilizes TensorFlow and TFLearn for building a 
3D Convolutional Neural Network (CNN) model 
that classifies whether a patient has lung cancer or 
not. The model is trained on CT scan images and 
uses softmax classification for prediction. The GUI, 
built using Tkinter, provides an interactive interface 
for importing data, preprocessing it, training the 
model, and displaying results, including accuracy, 
confusion matrix, and predictions. The system 
enables efficient and automated lung cancer 
diagnosis, assisting medical professionals in early 
detection and decision-making.