Related Experiment Video
Updated: May 20, 2025
07:53
Author Spotlight: Advancing 3D Modeling for Enhanced Diagnosis and Treatment of Pulmonary Nodules in Early-Stage Lung Cancer
Published on: October 13, 2023
1.3K
EMI-LTI: An enhanced integrated model for lung tumor identification using Gabor filter and ROI
Jayapradha J1,2, Su-Cheng Haw2, Naveen Palanichamy2
1Department of Computing Technologies, School of Computing, SRM Institute of Science and Technology, Kattankulathur, Tamil Nadu, 603203, India.
Methodsx
|March 24, 2025
Summary
This study enhances early lung cancer diagnosis using CT scans. The Enhanced Integrated model for Lung Tumor Identification (EIM-LTI) shows improved accuracy over Convolutional Neural Networks (CNN) in identifying cancerous regions.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oncology
Background:
- Early detection of lung cancer significantly improves patient outcomes.
- Accurate analysis of CT scans is crucial for timely diagnosis.
- Developing robust automated systems can aid radiologists in lung cancer detection.
Purpose of the Study:
- To analyze CT scan images for early-stage lung cancer diagnosis.
- To compare the performance of a Convolutional Neural Network (CNN) with a proposed Enhanced Integrated model for Lung Tumor Identification (EIM-LTI).
- To evaluate the effectiveness of image pre-processing techniques and data augmentation in lung cancer detection.
Main Methods:
- CT scan images of lung cancer patients were pre-processed using Gabor filters, contouring for Region of Interest (ROI) labeling, sharpening, and cropping.
- Data augmentation was applied to pre-processed images using CNN and EIM-LTI architectures.
- Model performance was evaluated using precision, sensitivity, F1-score, specificity, training/validation accuracy, and validation loss.
Main Results:
- The EIM-LTI model demonstrated higher training accuracy (2.67%) and validation accuracy (2.7%) compared to the CNN model.
- The EIM-LTI model had a slightly higher validation loss (0.0333) than the CNN.
- Cross-validation (5 folds) achieved 98.27% accuracy, with 92% accuracy on unseen data.
Conclusions:
- The EIM-LTI model offers a promising approach for enhanced early-stage lung cancer detection from CT scans.
- Image pre-processing and data augmentation are vital for improving the performance of deep learning models in medical diagnostics.
- The proposed EIM-LTI model shows superior performance in accuracy and sensitivity for lung cancer identification.
