2020 will be remembered as the year the novel coronavirus (COVID-19) swept across the world and impacted our public health systems, as well as our daily lives and work.
The virus has infected huge populations of people in a short span of time, which imposes huge pressure on the medical systems of many regions. As a result, the medical industry is looking for ways to improve and expedite diagnosis and treatment. Using the AI-empowered medical technology can contain the virus and save more lives.
The medical diagnosis and evaluation using CT images are important to preventing and controlling the spread of the COVID-19. At the peak of the pandemic, the number of patients in hospital every day usually surpasses the maximum treatment capacity of the hospital. This means that not only are most departments overwhelmed, but this is hindered by traditional CT technology, which is not capable of providing fast and accurate diagnosis. Specifically, the system processes image browsing, compares historical images, and describes impact before generating an analysis report for radiologists. In normal settings, conventional diagnosis technology takes 15–20 minutes and delivers an accuracy rate up to 90%. However, due to the limited medical resources and heavy workload of doctors, this figure may decline in practice. Consequently, this impacts the future treatment of suspected cases, false negatives, and confirmed cases.
At this critical juncture, Huawei works with medical partners to find a way forward. In March 2020, Huawei and Advanced Global Solution (AGS), an Italian-based enterprise, jointly launched an AI-based medical image diagnosis system to supercharge the diagnosis and treatment for COVID-19. This diagnosis solution runs on the AI foundation and ultimate computing power of the Atlas 800 server and Atlas 300 inference card needed in a wide range of intelligent functions, such as intelligent CT image reading, automatic pneumonia classification, multi-color rendering of lesions, and intuitive comparison of baseline histograms. Combined with the AI image diagnosis system and the expertise of doctors, the solution quickly identifies the symptoms of patients with COVID-19 within 2 to 3 minutes, and evaluates the treatment effect with an accuracy rate over 98%. The fast and accurate diagnosis is crucial to the treatment of pandemics like the COVID-19. With the Atlas-based intelligent solution, doctors can determine the treatment scheme as soon as possible and allocate medical resources based on the severity of the disease, providing effective and timely treatment for patients in critical condition.
The open Atlas 800 inference server is an easy-to-use device that offers high performance and adaptive configuration in a secure and trusted manner. It runs on the Kunpeng- and Intel-based AI inference platforms, and can be easily deployed within half a day, allowing hospitals to immediately provide AI-based diagnosis and treatment amid the medical emergency. This speed is so impressive that this solution has been widely adopted in many hospitals across the world, such as Italy, Poland, France, and Mexico, to deal with the growing number of patients.
Huawei focuses on open hardware, open source software, and partner enablement. Huawei is committed to providing full-stack, all-scenario AI computing infrastructure and application solutions to empower industries with AI and bring pervasive intelligence.