ViT-Lung: A Vision Transformer Framework for Lung Cancer Detection

View at publisher Summarise with VIRAT
Authors: B. Rajalakshmi, V. Murali Bhaskaran
Cited by: 0
References: 14

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

Lung cancer continues to be a major cause of death from cancer globally, primarily because it's often found at a late stage. Traditional imaging methods like CT scans and X-rays, which rely heavily on the skills of radiologists, have limitations. While deep learning, especially convolutional neural networks, has helped with automated detection, these methods often miss the bigger picture and struggle to identify early signs of the disease. To overcome this, a new system called VIT-LUNG, based on Vision Transformers, is being suggested for lung cancer detection. VIT-LUNG uses self-attention to understand how different parts of the image relate to each other, allowing it to more accurately spot subtle cancerous nodules. The system involves preparing the images through processes like adjusting brightness and reducing noise, followed by breaking the image into smaller pieces and analysing them with a transformer model. Techniques like data modification and using knowledge gained from other studies are also used to improve the system's performance and prevent it from being too specific to the training data. Tests on a widely used dataset showed that VIT-LUNG performed better than existing CNN-based methods in terms of accuracy and other important measures. In addition, the system generates visual representations highlighting the key areas of the image that influenced its predictions, making it easier to understand how it arrives at its conclusions. These results suggest that VIT-LUNG is a promising tool for early and accurate lung cancer detection. Its automated nature and ability to explain its reasoning could help doctors make better decisions and improve patient outcomes.

Fields of Study & Subjects

AI in cancer detection COVID-19 diagnosis using AI Lung Cancer Diagnosis and Treatment

Publication details

Journal / source
International Conference on Communication, Computing and Electronics Systems (ICCCES)
Pages
1909–1915
Publication date
21 January 2026
Language
EN