Outsmarting COVID-19


    Mar 20, 2020

    On March 19, China reported no new domestic transmissions for the first time since the coronavirus outbreak started. However, in much of the rest of the world the virus is continuing to spread, with nations grappling with the best way to halt the pandemic.   

    On the healthcare front, a fast and accurate diagnosis that enables rapid individual treatment and quarantine, while helping to alleviate the crushing workloads of healthcare professionals, is crucial to combating the impact of the virus.

    The Limits of Conventional Diagnosis

    To more accurately and quickly diagnose COVID-19, China’s national health commission included CT (computer tomography) in the clinical diagnosis standard for Hubei province on February 4, 2020.

    CT scans are a medical imaging procedure in which a computer processes combinations of many X-ray measurements taken from different angles. It then produces cross-sectional images of specific areas of a scanned object, allowing the user to see inside the object without cutting into it.

    The Drawbacks

    As one of the determinants for the diagnosis and treatment of COVID-19, CT scans are fast and accurate. However, due to the large numbers of lesions in the lungs and the rapid changes the virus creates, multiple rechecks and image reviews are required over a short period of time, significantly increasing doctors’ workloads.

    At the same time, there are too few imaging doctors who can accurately diagnose and quantitatively analyze COVID-19, resulting in low diagnosis efficiency.

    To develop an AI-assisted quantitative medical image analysis service specifically for COVID-19, HUAWEI CLOUD teamed up with Huazhong University of Science & Technology and Lanwon Technology. The solution applies computer vision and medical imaging analysis to quickly and accurately output CT quantification results for clinical doctors. It can help overcome the shortage of imaging experts who can accurately diagnose COVID-19, relieve the pressure of quarantining non-infected patients, and reduce the heavy workloads of doctors.

    How It Works

    The solution uses computer vision and medical image analysis to segment multiple pulmonary ground glass opacities (GGOs) and lung consolidation instances. GGOs are attenuations that decrease the amount of air in the lung. Lung consolidation occurs when air in the lung’s alveoli is replaced by liquid or other substances. The AI delivers quantitative evaluations of both conditions through CTs of patients’ lungs.

    UIs of the HUAWEI CLOUD NCP-CT AI-assisted Quantitative Diagnosis Service

    Having analyzed hundreds of infected and non-infected patients, the diagnosis service uses industry-leading Dice Similarity Coefficient (DICE) and Absolute Volume Difference (AVD). DICE and AVD deal with lesion area segmentation: DICE overlaps predicted lesions and actual lesions and AVD identifies volume differences between them. The results are consistent with precise manual sketching by doctors.

    The service is far more diagnostically efficient than conventional methods whereby doctors manually draw Regions of Interest (ROIs) for quantitative evaluation. The computing capabilities of Huawei Ascend series AI chips can output the quantization result of a single case in seconds, making AI coupled with doctor review dozens of times faster than manual quantitative image evaluation.

    By combining clinical information and laboratory results, doctors can more accurately distinguish between early, advanced, and severe stages of COVID-19, facilitating early screening and prevention/control. For confirmed cases in hospitals, this AI-assisted service can quickly perform registration and quantitative analysis on the 4D dynamic data of multiple rechecks, helping doctors evaluate patients’ conditions and the effect of drug treatment.

    AI-assisted Quantitative Analysis On 4D Dynamic Data

    The AI + CT medical image analysis service for COVID-19 is based on the HUAWEI CLOUD EIHealth-medical image analysis platform, which uses the AI Ascend cluster service and ModelArts one-stop AI development and management platform to implement:

    • One-stop medical image data governance and labeling
    • Model training and evaluation
    • Visualized rendering

    These functions provide powerful support for universities and hospitals by delivering massive AI computing, platform, and algorithm power.

    Recently, HUAWEI CLOUD has been working with various industries to carry out multiple AI-based scientific research projects on COVID-19, releasing the results of ultra-large-scale computer-assisted drug screening with many scientific research institutions. 

    Click the links to read more about HUAWEI CLOUD and how ICT can help to mitigate the effects of the pandemic on life and work. Subscribe to this blog for more updates on ICT’s application to the coronavirus pandemic as well as the latest tech trends.

    Post adapted from this press release on huaweicloud.com

    Disclaimer: Any views and/or opinions expressed in this post by individual authors or contributors are their personal views and/or opinions and do not necessarily reflect the views and/or opinions of Huawei Technologies.


      Leave a Comment

      Posted in


      Posted in