Classification of Non-Small Cell Lung Cancer Using Deep Learning
Computers can be programmed using Deep learning, a subset of Artificial Intelligence that takes inspiration from the human brain. This approach is quicker and less tedious. The major drawback in the bio medical field is that fewer datasets are available which can lead to overfitting which in turn re...
Saved in:
Published in | 2023 International Conference on Applied Intelligence and Sustainable Computing (ICAISC) pp. 1 - 5 |
---|---|
Main Authors | , , , , |
Format | Conference Proceeding |
Language | English |
Published |
IEEE
16.06.2023
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | Computers can be programmed using Deep learning, a subset of Artificial Intelligence that takes inspiration from the human brain. This approach is quicker and less tedious. The major drawback in the bio medical field is that fewer datasets are available which can lead to overfitting which in turn reduces the accuracy. This can be overcome with the help of data augmentation. The adenocarcinoma (ADC), squamous cell carcinoma (SCC), and large cell carcinoma (LCC), subtypes of Non-Small Cell Lung Cancer are classified using one of the Deep Learning techniques in this study. This is accomplished by utilizing the EfficientNetB2 architecture which consists of 342 layers with additional neurons on top it. The dataset is made up of 1000 CT (Computed Tomography) scan images of lungs, which has been broken down into training (613 images), testing (72 images), and validation (315 images) sets. The training accuracy of 95% and testing accuracy of 83% is obtained. |
---|---|
DOI: | 10.1109/ICAISC58445.2023.10200012 |