As of the end of 2018, artificial intelligence in the field of medicine is still a doctor-led situation, assisted by artificial intelligence.
Artificial Intelligence for the field of chronic disease breakthroughs:
Major technology companies have put their attention to the field of chronic disease, with the development of artificial intelligence, the prediction and early diagnosis of chronic disease has now gained significant improvement.
Tencent's Parkinson's AI-assisted diagnostic technology, based on motion video analysis technology, can automatically realize UPDRS (the international universally adopted Parkinson's Disease Rating Scale) scoring based on the motion video of Parkinson's patients, and with the assistance of the AI technology, the user does not need to wear any sensors, and only needs to be photographed through the camera (which can be satisfied by an ordinary smartphone) to realize daily assessment of the motor function of Parkinson's disease, and the doctor can realize the daily assessment of Parkinson's disease.
The AI technology can be used in the daily assessment of Parkinson's disease, and the doctor can complete the diagnosis process within 3 minutes, which is 10 times faster than the diagnosis.
Ali launched the "Ruining sugar", through a large number of doctors' practical experience as an empirical model, a large number of medical knowledge and authoritative literature as a knowledge model, the use of a series of Internet of Things management, the use of artificial intelligence fundus lesions and urine protein screening technology, based on deep computer learning to establish the Diabetes and complications screening software, to realize the "artificial intelligence" of diabetes patients from prevention, diagnosis, treatment, to complications management.
Meanwhile, Korean researchers have used a database of brain images of healthy people and Alzheimer's disease patients created by Alzheimer's disease researchers around the world to train a convolutional neural network that recognizes the differences between them. The software system identified patients with mild cognitive impairment patients transforming into Alzheimer's disease with a predictive accuracy of 84.2%, outperforming conventional feature-based human quantification methods and showing the feasibility of deep learning techniques to predict disease prognosis using brain images.