Sunil Kumar, Amar Pal Yadav, Neha Nandal, Vishal Awasthi, Luxmi Sapra, Prachi Chhabra

COV-TViT: An Improved Diagnostic System for COVID Pneumonitis Utilizing Transfer Learning and Vision Transformer on X-Ray Images

Introduction

Cov-tvit: an improved diagnostic system for covid pneumonitis utilizing transfer learning and vision transformer on x-ray images . COV-TViT: An AI diagnostic system for COVID pneumonitis using X-rays. Integrates transfer learning with Vision Transformer for 98.96% accurate detection, outperforming CNNs for efficient medical imaging.

Abstract

COVID is a contagious lung ailment that continues to be a world curse, and it remains a highly infectious respiratory disease with global health implications. Traditional diagnostic methods, such as RT-PCR, though widely used, are often constrained by high costs, limited accessibility, and delayed results. In contrast, radiology for lung disease detection has been proven advantageous for identifying deformities, and chest X-rays are the most preferred radiological method due to their non-invasive nature. To address these limitations, this study aims to develop an efficient, automated diagnostic system leveraging radiological imaging, specifically X-rays, which are cost-effective and widely available. The primary contribution of this research is the introduction of COV-TViT, a novel deep learning framework that integrates transfer learning with Vision Transformer (ViT) architecture for the accurate detection of COVID pneumonitis. The proposed method is evaluated using the COVID-QU-Ex dataset, which comprises a balanced set of X-ray images from COVID positive and healthy individuals. Methodologically, the system employs pre-trained convolutional neural networks (CNNs), specifically VGG16 and VGG19 (Visual Geometry Group), for transfer learning, followed by fine tuning to enhance feature extraction. The ViT model, known for its self-attention mechanism, is then applied to capture complex spatial dependencies in the X-ray images, enabling robust classification. Experimental results demonstrate that COV-TViT achieves a classification accuracy of 98.96% and an F1 score of 96.21%, outperforming traditional CNN based transfer learning models in several scenarios. These findings underscore the model’s potential for high-precision COVID pneumonitis detection. The proposed approach significantly transforms classification tasks using self-attention mechanisms to extract features and learn representations. Overall, the proposed diagnostic system COV-TViT can be advantageous in the fundamental identification of COVID pneumonitis.

Review

The paper addresses a highly relevant and pressing global health issue: the rapid and accurate diagnosis of COVID pneumonitis. Acknowledging the inherent limitations of traditional diagnostic methods such as RT-PCR in terms of cost, accessibility, and turnaround time, the authors propose an innovative solution leveraging the widespread availability and cost-effectiveness of chest X-ray imaging. The core contribution of this research is COV-TViT, a novel deep learning framework designed to create an efficient and automated diagnostic system by integrating transfer learning with the powerful Vision Transformer (ViT) architecture for enhanced detection capabilities. Methodologically, the COV-TViT system employs a sophisticated two-step approach. It first utilizes pre-trained Convolutional Neural Networks (CNNs), specifically VGG16 and VGG19, for transfer learning to effectively extract initial features from the X-ray images, followed by fine-tuning to optimize this process. Crucially, a Vision Transformer (ViT) model is then applied, leveraging its self-attention mechanism to capture complex and long-range spatial dependencies within the X-ray data. This allows for a more comprehensive understanding of the radiological patterns associated with COVID pneumonitis, leading to robust classification. The proposed framework's performance is thoroughly evaluated using the balanced COVID-QU-Ex dataset, comprising X-ray images from both COVID-positive and healthy individuals. The experimental results presented are highly encouraging, with COV-TViT achieving a remarkable classification accuracy of 98.96% and an F1 score of 96.21%. These metrics demonstrate a significant improvement over traditional CNN-based transfer learning models, underscoring the efficacy of combining transfer learning with the Vision Transformer's self-attention capabilities for feature extraction and representation learning. The findings highlight the model's strong potential for high-precision COVID pneumonitis detection, suggesting that COV-TViT could indeed be a transformative and advantageous tool in the fundamental and rapid identification of this critical lung ailment in real-world clinical settings.

Full Text

Comments

You need to be logged in to post a comment.