A Medical AI Diagnosis Platform Based on Vision Transformer for Coronavirus

With the spread of the novel coronavirus in the world, COVID-19 has raised a number of serious issues regarding its diagnosis and treatment. Currently, most COVID-19 patients are diagnosed using a lung CT scan, which is extremely inefficient, especially when faced with a large number of patients. This manual method not only increases the workload for doctors, but it can also delay patient treatment. Thus, many scholars and companies have developed auxiliary diagnosis platforms to speed up the diagnosis process. However, due to limitations surrounding the development of imaging technology, the accuracy of current platform-assisted diagnosis is still relatively low. In order to solve these problems, this paper designs a medical AI dialogue diagnosis platform based on the Vision Transformer by distilling technology to gain more medical information from the traditional image recognition model and significantly improve the predictive performance of COVID-19. At the same time, this paper also proposes the addition of a simple dialogue system to improve the efficiency of man-machine interactions. With this, it can be concluded that the medical dialogue system for COVID-19 detection can realize its anticipated function and has certain practical significance.