Covid-AI detects COVID-19 infections in x-rays. It can differentiate between COVID-caused and other types of early-stage pneumonia. (Sensitivity to COVID: 99%, Accuracy: 97%.) The model was developed as a faster, cheaper detection method than RC-PCR screening. It is part of a broader program for detecting pathologies present in x-ray images.

Value Proposition

X-ray machines are widely available, and taking images is inexpensive. Detecting COVID in x-rays increases access to testing, speed (vs. PCR), and allows patients to maintain safe social distancing.

Narrative

A doctor captures a patient’s chest x-ray images. The patient consents to the doctor using those images for research. The doctor accepts the terms and conditions and uploads the x-ray images. He receives the screening results.

Intended Use

Bilateral chest x-rays from patients with pneumonia symptoms over the age of 18.

Limitations

Model performance metrics are based on publicly available datasets. It should be used for research purposes only. The creators are currently improving it based on user feedback.

Model performance metrics

Accuracy: 0.98
Sensitivity: 0.989
Specificity: 0.971
F1Score:0.979
NPV:0.99
PPV:0.96 

CovidAi

AI-HUT

AI-HUT is an advanced, world-class, and up-to-date AIs lab that uses the latest scientific and technical tools to bring value to the world. We are focusing on the healthcare and education sectors as we understand the importance of these sectors for enhancing and supporting the challenges facing society and industry.

With highly qualified team members worldwide and partners in the US, Canada, Armenia, India and Syria, UAE-headquartered AI-HUT takes the initiative to meet the needs of implementing advanced AI solutions efficiently.

We’re a startup that offers a wide range of products and services to individuals, companies, and societies who want to benefit from AI.

Ghaithpic

Ghaith Sankari

Learn more