In the ambulance, time with God Why is this short process enough to become an "AI good question" that many companies cannot find? Chen Dizhi, chief strategy officer of Shangzhizhi, pointed out that the abnormal ECG judgment before the ambulance arrives at the hospital actually has four characteristics of medical AI applications: 1. First, the abnormal ECG judgment can be carried out through AI in the ambulance, so that the patient can be sent to the most The right hospital and the most suitable surgery have the "immediacy" where every second counts. 2.
People's personal information will be "anonymized" to avoid privacy disputes. 3. Appropriate intervention of AI can "relieve the pressure on the scene". Fourth, physicians can use the status of "expert endorsement" to make AI the best assistant in the auxiliary medical field. DSC00661 Chen Dizhi, Chief Strategy Officer of Shangzhizhi Photo Credit: The News Lens Brand Studio Because of these four characteristics, when the electrocardiogram data whatsapp list is immediately transmitted from the ambulance to the platform, the data is connected through Amazon SageMaker, and it is immediately determined who is the AMI patient, and then AWS will immediately transmit the abnormal data to the communication group. Personnel can then prepare for subsequent rescue procedures in advance.
When the ambulance staff no longer have to wait for the doctor to judge from a distance, and the medical staff no longer have to bear excessive pressure and constantly check whether there is an emergency message from the mobile phone, the medical scene can make good use of the precious time grabbed from the hands of God and concentrate more on the patient. Prepare for follow-up surgery. Medical AI is not easy to implement, machine learning can help The reason why we can grab time from God is machine learning. Machine learning is good at solving problems that are "independent of the external context", that is, those problems that can be inferred only by judging a single or a few known kinds of data.