Huawei develops AI-based COVID-19 diagnosis platform
One of the major problems in cracking down the rapid spread of Covid-19 is that conducting testing for the virus can be extremely problematic – even when a patient has undergone a CT scan which is generally considered a fast and effective way to diagnose and evaluate a patient.
The main issue is that Covid-19 scans can involve a medical specialist reviewing a large numbers of lesions in a patients lungs and rapid ongoing changes in the patients system can mean a CT Scan requiring multiple rechecks and image reviews in a short period of time.
This represents a major barrier for a quick and accurate diagnosis because there are only a limited number of specialists normally present at a hospital who can review the patients CT scan – which inevitably slows down the ability to quickly diagnose patients.
With Covid-19 becoming a deepening public health crisis in China and around the world Huawei worked with the Huazhong University of Science & Technology and Lanwon Technology to develop and launch an AI-assisted quantitative medical image analysis service to accelerate the testing and diagnosis of COVID-19.
Using Huawei’s leading AI technologies such as computer vision and medical image analysis the service can automatically and correctly output CT quantification results to imaging and clinical doctors.
This rapidly speeds up the diagnostic process because using AI alleviates the on-the-ground shortage of imaging doctors who can accurately diagnose COVID-19.
The diagnostic platform utilizes the Huawei Ascend series AI chips to output the quantization result of a single case in seconds, this is dozens of times faster than manual quantitative image evaluation and also greatly improves the diagnosis efficiency.
In addition, the Huawei platform also leverages the computer vision and medical image analysis technologies to segment the multiple pulmonary ground glass opacities (CGOs) and lung consolidation and make quantitative evaluations through CTs of patients’ lungs.
It combines clinical information and laboratory results to help doctors more accurately distinguish between early, advanced, and severe stages of COVID-19 which helps to facilitate early screening and prevention/control.
For confirmed cases of Covid-19 in hospitals the technology can perform registration and quantitative analysis on the 4D dynamic data of multiple re-checks within a short period of time, helping doctors effectively evaluate patients conditions and drug use effects on an ongoing basis.
According to the analysis of hundreds of COVID-19 cases and non-infected cases, our technology has achieved an extremely high level of accuracy in tracing the predicted lesions present in a patient and the actual lesions present in the patients and are highly consistent with the precise manual sketching by doctors.
Once the highly accurate results have been recorded then the technology platform can produce CT quantization results in seconds – substantially faster than if a doctor were to try to produce the results manually, meaning that many more tests can be successfully conducted and outputted.
The medical image analysis service for COVID-19 is developed and constructed based on the Huawei e-Health-medical image analysis platform.
The platform uses the Huawei AI Ascend cluster service and the ModelArts one-stop AI development and management platform to implement one-stop medical image data governance and labeling, model training and evaluation, and visualized rendering, providing powerful support for universities and hospitals with massive AI computing, platform, and algorithm power.