Scholarlink

A dangerous virus spread through SARS (severe acute respiratory syndrome), mainly affecting the respiratory system, known as Coronavirus. At the beginning of this disease, there were no specific techniques to identify it due to the similarity of its symptoms with cold and cough. However, some immediate technologies were developed for detection because of its spreading effects and death rates. To predict the accuracy and faster results, many AI derivative models were developed on available datasets. In this paper, initially, a detailed survey on such techniques is provided. This detailed survey includes the technique and methodology used, model used, datasets taken, accuracy achieved, and related parameters. Later on, we proposed an efficient approach using a Transfer Deep Learning approach based on public datasets composed of lungs and chest Computer Tomography images, X-ray images, and ultrasound images of the Chest and lungs. The proposed model is tested on various parameters like TNR, ERR, Accuracy, TPR, FPR, F1-Score, Error Rate, and FNR based on a confusion matrix. Our method achieved 95% accuracy, 96% TPR, 4% ERR, and TNR of 95%. After detection, we can start the appropriate treatment of the patient.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Detection of Severe Acute Respiratory Syndrome Corona Virus Disease Using Transfer Learning Approach

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Krishna Kumar Joshi,
  • Kamlesh Gupta,
  • Jitendra Agrawal

摘要

A dangerous virus spread through SARS (severe acute respiratory syndrome), mainly affecting the respiratory system, known as Coronavirus. At the beginning of this disease, there were no specific techniques to identify it due to the similarity of its symptoms with cold and cough. However, some immediate technologies were developed for detection because of its spreading effects and death rates. To predict the accuracy and faster results, many AI derivative models were developed on available datasets. In this paper, initially, a detailed survey on such techniques is provided. This detailed survey includes the technique and methodology used, model used, datasets taken, accuracy achieved, and related parameters. Later on, we proposed an efficient approach using a Transfer Deep Learning approach based on public datasets composed of lungs and chest Computer Tomography images, X-ray images, and ultrasound images of the Chest and lungs. The proposed model is tested on various parameters like TNR, ERR, Accuracy, TPR, FPR, F1-Score, Error Rate, and FNR based on a confusion matrix. Our method achieved 95% accuracy, 96% TPR, 4% ERR, and TNR of 95%. After detection, we can start the appropriate treatment of the patient.