The world has been facing the COVID-19 pandemic since December 2019. Timely and efficient diagnosis of COVID-19 suspected patients plays a significant role in medical treatment. The deep transfer learning-based automated COVID-19 diagnosis on chest X-ray is required to counter the COVID-19 outbreak. This work proposes a real-time Internet of Things (IoT) framework for early diagnosis of suspected COVID-19 patients by using ensemble deep transfer learning. The proposed framework offers real-time communication and diagnosis of COVID-19 suspected cases. The proposed IoT framework ensembles four deep learning models such as InceptionResNetV2, ResNet152V2, VGG16, and DenseNet201. The medical sensors are utilized to obtain the chest X-ray modalities and diagnose the infection by using the deep ensemble model stored on the cloud server. The proposed deep ensemble model is compared with six well-known transfer learning models over the chest X-ray dataset. Comparative analysis revealed that the proposed model can help radiologists to efficiently and timely diagnose the COVID-19 suspected patients.
Download full-text PDF |
Source |
---|---|
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8197673 | PMC |
http://dx.doi.org/10.1155/2021/3277988 | DOI Listing |
Cureus
December 2024
Department of Internal Medicine, Hiroshima City Funairi Citizens Hospital, Hiroshima, JPN.
Although human metapneumovirus(hMPV) infection can induce severe symptoms in older adults or immunocompromised patients, it usually causes mild symptoms in young immunocompetent adults. The prevalence of hMPV infectious disease is highest during the late winter and early summer. We report a hypoxemic case of hMPV infection in a young immunocompetent man that occurred in the first autumn after the reclassification of coronavirus disease (COVID-19) from Class 2 to Class 5.
View Article and Find Full Text PDFIJID Reg
March 2025
Postgraduate Program in Parasitic Biology, Federal University of Sergipe, Sergipe, Brazil.
Objectives: To investigate the prevalence of nine respiratory viruses and their clinical characteristics in children aged up to 5 years old in the state of Sergipe, Northeast of Brazil in the pre-COVID-19 pandemic period.
Methods: Children with suspected influenza virus infection were included in the study. Clinical samples were screened using real-time quantitative polymerase chain reaction for the diagnosis of adenovirus, parainfluenza (PIV)1, PIV2, PIV3, and human metapneumovirus.
Acta Med Indones
October 2024
Department of Cardiology and Vascular Medicine, Faculty of Medicine Universitas Indonesia - Harapan Kita National Cardiovascular Center, Jakarta, Indonesia.
Increased thrombotic events that occur in up to one-third of patients with COVID-19 are predominantly pulmonary emboli (PE), which are associated with higher severity and increased mortality. Acute PE should therefore be one of the main differential diagnoses among patients with hemodynamic instability. Early treatment of such a condition with systemic thrombolysis remains the first line of treatment especially in patients with COVID-19, which hinders further invasive intervention.
View Article and Find Full Text PDFBMJ Open Qual
January 2025
Professor Department of Obstetrics and Gynaecology, Lady Hardinge Medical College, New Delhi, India.
Background: Allowing a birth companion is the basic right of a mother and is identified as an important component of respectful maternity care. The implementation of this intervention has been a challenge in heavy-load public health facilities in India.
Local Problem: Despite the proven benefits of the presence of birth companions on maternal-fetal outcomes, there was no policy of allowing birth companions in our hospital.
J Infect Public Health
January 2025
Division of Clinical Pathology, Department of Pathology, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan. Electronic address:
Background: We aimed to evaluate the efficacy of integrating the Varia5 multiplex assay (qPCR) and whole genome sequencing (WGS) for monitoring SARS-CoV-2, focusing on their overall performance in identifying various virus variants.
Methods: This study included 140 naso-pharyngeal swab samples from individuals with suspected COVID-19. We utilized our self-developed Varia5 multiplex assay, which targets five viral genes linked to COVID-19 mutations, in conjunction with comprehensive genomic analysis performed through whole genome sequencing (WGS) using the Oxford Nanopore system.
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!