Research Article
Deep Learning Approaches for Precise Lung Cancer Diagnosis
@INPROCEEDINGS{10.4108/eai.28-4-2025.2357792, author={M. Sreedevi and Mudiveti Harshitha Reddy and Pasala Amrutha and Endluri Deepthi and Mekala Pavani}, title={Deep Learning Approaches for Precise Lung Cancer Diagnosis}, proceedings={Proceedings of the 4th International Conference on Information Technology, Civil Innovation, Science, and Management, ICITSM 2025, 28-29 April 2025, Tiruchengode, Tamil Nadu, India, Part I}, publisher={EAI}, proceedings_a={ICITSM PART I}, year={2025}, month={10}, keywords={image preprocessing feature extraction \& segmentation vgg19 efficientnet-b0 resnet50}, doi={10.4108/eai.28-4-2025.2357792} }- M. Sreedevi
Mudiveti Harshitha Reddy
Pasala Amrutha
Endluri Deepthi
Mekala Pavani
Year: 2025
Deep Learning Approaches for Precise Lung Cancer Diagnosis
ICITSM PART I
EAI
DOI: 10.4108/eai.28-4-2025.2357792
Abstract
The leading source of lung cancer serving as a significant health issue among men and women stems from tobacco contact and smoking habits. Medical technology developments since modern times have failed to decrease the substantial mortality from this disease. Early detection of lung cancer relies on successful diagnosis through machine learning technology that has evolved into the best diagnostic method for the disease. The achievement of precise medical diagnosis using different classification systems remains a difficult task due to elusive perfection levels. The poor management of Digital Imaging and Communications in Medicine (DICOM) images creates additional costs in implementation. Medical imaging professionals primarily use CT scanners because this equipment generates clearer images with lower noise levels during the scanning procedure. Deep learning has achieved two main objectives: the detection of nodules and the assessment of abnormal structures and cancer progression in respiratory tissue. The ability to detect conditions has seen major improvements resulting from models utilizing EfficientNet-B0, VGG19 and ResNet-50. The most efficient model for delivering top results is the ResNet-50 model among the available choices.
