| Scenario | Technical feature | Typical case |
| Medical imaging and disease diagnosis | Computer vision (lesion detection, image segmentation); NLP generates structured reports | The accuracy rate of identifying COVID-19 lung lesions is as high as over 90%; Google DeepMind's breast cancer screening AI is more accurate than radiologists; Eagle Eye Technology's diabetic retinopathy screening system; Low-dose CT scans have a sensitivity of 84%-90% in detecting lung nodules |
| Drug research and development | Machine learning accelerates target discovery and compound screening; NLP analyzes literature and experimental data | Insilico Medicine uses generative AI to design new drug molecules, shortening the R&D cycle to one year |
| Medical robot | Multimodal perception of surgical robots (vision+force feedback); dynamic adjustment scheme of rehabilitation robots | Minimally invasive robot Tumai® laparoscopic surgical robot approved by NMPA; 5G remote orthopedic robot surgery at Beijing Jishuitan Hospital |
| AI-optimized clinical trials | Optimize patient recruitment plans, reduce recruitment time and costs; assist in determining appropriate clinical endpoints | Cincinnati Children's Hospital Medical Center uses machine learning to analyze why people accept or reject invitations to participate in clinical trials, reducing subject screening time by 34% |
