At the epicenter of the Coronavirus in Wuhan China, a team of physicians in China are using GPU-accelerated AI software to detect visual signs of the coronavirus (Covid 19).
Physicians there say the AI-based software, which relies on NVIDIA GPUs for both training and inference, has helped overworked staff screen patients and prioritize those likely to have the virus.
The software, developed by Beijing-based startup Infervision, a member of the NVIDIA Inception, NVIDIA’s startup accelerator program, can identify typical signs or partial signs of Covid 19 infection. To do this, Infervision’s software detects signs of pneumonia caused by the virus.
Once the outbreak began, the company quickly fine-tuned their existing pneumonia models to look more closely at coronavirus pneumonia.
Training was performed on NVIDIA V100 GPUs using more than 2,000 CT images from some of the first coronavirus patients in China. In terms of development software, the company relies on NVIDIA Clara SDKs, NVIDIA’s AI healthcare application framework for AI-powered Medical Imaging.
One of the major advantages of the AI-system is speed. More conventional methods can take longer to process due to the increase in demand. However, with the software, doctors can quickly determine the best treatment options.
Before this outbreak, the software was primarily used to detect cancer in lung CTs.
So far, the new end-to-end system has already been deployed at 34 hospitals in China and used to review more than 32,000 cases of the coronavirus or Covid 19.
As an end-to-end system, Infervision models use NVIDIA T4 GPUs for inference.
Infervision systems have been trained on hundreds of thousands of lung images collected from hospitals in China. In addition to being used in China, Infervision’s software is currently in evaluation by clinics in Europe and the United States.
“Infervision is fully committed to support the medical professionals and stand with the patients in controlling the 2019-nCoV outbreak,” the company stated in a LinkedIn post.