X-MedVIT: Interpretable Vision Transformers for Reliable Pneumonia Detection in Medical Imaging

Read full text Summarise with VIRAT
Authors: Jayaraj R, Kishore Kumar G B, Praveen M, Lokesh M
Cited by: 0
References: 5

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

AI has become an important tool in medical diagnostic systems. Deep learning models often lack transparency that makes it difficult for clinics to use them. To address this, we propose X-MedVIT, an interpretablemedical imaging framework based on Vision Transformers (ViT). This system analyzes the chest X-ray images and issues a diagnosis and an explanation of the rationale behind the decision. It incorporates attention rollout, gradient times input saliency, and token-level relevance propagation. Unlike standard ViT models that do not take into account the generalisation performance, X - MedVIT also provides visual and content-related evidence that suggests which areas influenced the prediction. Our experimental public datasets we process shows an average accuracy of approximately 95%, demonstrating the reliable performance and boosting clinical trust and reasoning transparency. By offering predictions The system allows for digital health care solutions and helps to support global targets such as the United Nation Sustainable Development Goal, in order to improve health and access to diagnostics. Keywords— Vision Transformers, Explainable AI, Medical Diagnosis, Chest X-ray Analysis, Clinical Transparency.

Fields of Study & Subjects

COVID-19 diagnosis using AI Explainable Artificial Intelligence (XAI) Artificial Intelligence in Healthcare and Education

Publication details

Journal / source
INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT
Volume
10
Issue
03
Pages
1–9
Publication date
31 March 2026
Language
EN