Lung Cancer Stage Identification Using Enhanced Model From Deep Learning

Abstract

Ddigital image processing techniques such as classification and segmentation are now widely applied in the medical area with the aim of early detection of diseases. From the lungs Computed Tomography (CT) scan images the images are pre-processing and the Region of Interest (ROI) is segmented. In this work trying to develop CNN algorithm for Lung Cancer detection from CT-SCAN images and for CNN training. We use CT-SCAN images dataset which is downloaded from Kaggle.com website. This application uses hybrids algorithm of CNN-LSTM to enhance the performance of lung cancer prediction. The first step is to build a hybrid model using 80% of the x-ray images of lung cancer and the second step is to use the remaining 20% of the data to classify lung cancer. Present in the CT scan are of two types’ Normal’ and ‘ABNORMAL’. If the combined method (LSTM and CNN) will be used, accuracy, precision, fscore and recall will be developed far better than CNN. CNN algorithm is obtained 97% accuracy while hybrid CNN and LSTM has above 98% accuracy which is superior to CNN algorithm.
Keywords: Deep Learning (DL), Convolutional Neural Network (CNN), Hybrid Algorithm, Image Processing, Stage Identification, Medical Imaging.