Performance Evaluation and Optimization of Pre-trained Deep Learning Models Using a Weighted Ensemble Approach for Lung Cancer Classification

DOI

10.2991/978-94-6239-678-4_6How to use a DOI?

Keywords

Lung Cancer Detection; VGG16; Histopathological Image; ResNet50; DensNet121; Ensemble Learning

Abstract

Lung cancer remains to be among the leading contributor of deaths caused by cancer in the world, and accurate early diagnosis is a key to the better survival of the patient. The recent trends have observed the deep learning models as a promising tool in medical image analysis that can be used to accurately and automatically classify cancer. The paper presents the comparison of experimental study of three pre-trained convolutional neural network (CNN) models, namely VGG16, ResNet50 and DenseNet121, with the use of publicly available LC25000 lung and colon histopathological image dataset. The best architecture of lung cancer detection was evaluated based on the models in terms of accuracy, precision, recall, and loss measures. The experimental findings indicate that VGG16 was the best in this case because it achieved the best validation accuracy of 99.88 percent and the lowest loss rates. The validation accuracy of DenseNet121 and ResNet50 were 99.52 and 97.84 respectively, neither being quite as competitive as VGG16 but still quite competitive by comparison. Besides that, a weighted ensemble model was also constructed by integrating all the three networks using weights of 0.3 (VGG16), 0.2 (ResNet50), and 0.5 (DenseNet121). The accuracy of the ensemble as a whole was 96.88% which, though good, was not able to beat the best individual model. The results show that despite the ability of ensemble technique to increase robustness, individual models, in particular VGG16, may perform better in the classification of histopathological images. The results point to the promise of deep learning for accurate lung cancer detection and underscore the need for comparative model analysis and optimized ensemble approaches to promoting computer-aided diagnostic technology in medical practice.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Neha Raja Panwar
AU  - Prashant Kumar Shrivastava
PY  - 2026
DA  - 2026/05/28
TI  - Performance Evaluation and Optimization of Pre-trained Deep Learning Models Using a Weighted Ensemble Approach for Lung Cancer Classification
BT  - Proceedings of the 2nd International Conference on Recent Advancement and Modernization in Sustainable Intelligent Technologies & Applications (RAMSITA-2026)
PB  - Atlantis Press
SP  - 58
EP  - 72
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-678-4_6
DO  - 10.2991/978-94-6239-678-4_6
ID  - Panwar2026
ER  -