Multi-Class Pneumonia Classification using Swin Transformer on Chest X-ray Images

View at publisher Summarise with VIRAT
Authors: Pratham Kaushik, Vinay Kukreja
Cited by: 0
References: 17

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This study discusses the use of a Swin Transformer model for the classification of multiple types of Pneumonia using chest X-ray images. The primary aim was to achieve an effective deep learning model that was able to distinguish between four classes: i.e., Covid-19, Normal, Pneumonia-Bacterial, and Pneumonia-Viral. The dataset was curated from publicly available sources and preprocessed by resizing, CLAHE, and ROI extraction to enhance the visibility of the lungs and reduce the features that are irrelevant. The model has been trained using the AdamW optimizer, learning rate schedule, and early stopping to prevent overfitting. The final model gave 9 8. 4 1% accuracy on training and a total accuracy of 94% with a validation accuracy of 92.59%. Performance was assessed with various measures such as precision, recall, F1 score, and confusion matrix, which were found to be excellent for the Normal and Pneumonia-Bacterial classes. Despite good performance, the model had difficulty discriminating between the two diseases - COVID-19 and Pneumonia Virus. Future work will focus on improving the generalization of models, looking into bigger datasets, and improving fine-grained classification techniques.

Fields of Study & Subjects

COVID-19 diagnosis using AI Radiomics and Machine Learning in Medical Imaging Pneumonia and Respiratory Infections

Publication details

Journal / source
International Conference on Communication, Computing and Electronics Systems (ICCCES)
Pages
1866–1870
Publication date
21 January 2026
Language
EN