Clinical Informatics System to Understand Lung Cancer Using Transfer Learning-Based Computer-Aided Application

Document Type

Conference Proceeding

Source of Publication

2023 7th International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT)

Publication Date

10-28-2023

Abstract

Lung cancer is the second leading cause of death worldwide after cardiovascular disease. Due to late diagnosis, lung cancer patients have an extremely low survival rate when compared to other cancer patients. Hence, early lung cancer detection is critical for patients to undergo early therapies, boosting their chances of surviving or potentially becoming cancer-free. Over the past few years, researchers have developed several deep learning-based computer-aided diagnostic tools to detect lung cancer early. As a consequence, deep learning models are readily susceptible to over-fitting for fewer data samples and can be computationally time-consuming to pick hyper-parameters, which usually leads to a reduction in performance. Our proposed technique came up with a solution to this challenge of lung cancer classification tasks that used a layer-wise transfer learning strategy. In this work, three types of data samples such as Lung adenocarcinoma (Lung a), lung squamous cell carcinoma (Lung scc), and lung benign (Lung b) have been taken and implemented models, namely Group-I, II, III with frozen layers and baseline algorithms as VGG-19. To check the validity of the proposed models on lung samples and attain the best classification accuracy on Lung scc vs. Lung b 97.53%, also discriminate between Lung a vs. Lung b testing accuracy 92.59%. In contrast, the remaining Lung scc vs. Lung a attained 89.30%. Finally, a comparison to other studies demonstrating the suggested model superiority over state-of-the-art models in testing accuracy

ISBN

979-8-3503-4215-4

Publisher

IEEE

Volume

00

First Page

1

Last Page

6

Disciplines

Computer Sciences | Medicine and Health Sciences

Keywords

Training, Computational modeling, Transfer learning, Lung, Lung cancer, Predictive models, Data models

Indexed in Scopus

no

Open Access

no

Link to Full Text

Share

COinS