Mohammad Nabiel Dwi Ananda, Jumanto Unjung
The rapid advancement of technology has led to significant developments in various fields, especially in the medical sector. Technological innovations in the medical field are essential to assist medical professionals and doctors in diagnosing diseases quickly and accurately, particularly for patients requiring urgent care. According to data released by the World Health Organization (WHO), lung cancer is one of the most frequently diagnosed and deadliest cancers worldwide. Many patients struggle to receive timely treatment due to delayed medical intervention and limited diagnostic tools. This study aims to develop an automated lung cancer detection system using a deep learning approach with the ResNet101 model and a fine-tuning strategy to improve classification accuracy. The dataset used in this research is the IQOTH/NCCD - Lung Cancer Dataset sourced from Kaggle, consisting of X-ray images classified into three types of lung cancer. The initial training process utilizes pretrained weights from ImageNet, followed by fine-tuning certain layers to enhance the model's generalization on new data. The novelty of this study lies in the integration of tailored data augmentation and class imbalance handling strategies with the fine-tuned ResNet101 architecture, coupled with systematic hyperparameter optimization, which collectively differentiate it from prior works. The experimental results show that the proposed model achieves an accuracy score of 98.08%, an improvement from the initial 95%. The model also demonstrates stable performance across other evaluation metrics such as recall, precision, and F1-score. Therefore, this system has the potential to serve as an effective and efficient early diagnostic tool for lung cancer, providing support in clinical decision-making for medical practitioners. © 2025 IEEE.
Universitas Negeri Semarang, Department of Computer Science, Semarang, Indonesia
