Home / SmartTech / Medical ultrasound / AI imaging, supportive exoskeletons and neuronal weather modeling – TechCrunch

Medical ultrasound / AI imaging, supportive exoskeletons and neuronal weather modeling – TechCrunch



In the time of COVID-19, much that is transferred from the scientific world to the general public understandably relates to the virus. But other areas are still active in medical research – and, as usual, there are countless interesting (and encouraging) stories that shouldn't be lost in the angry activity of coronavirus reporting. This past week brought good news for various ailments, as well as some innovations that could improve weather reporting and potentially save a few lives in Cambodia.

Ultrasound and AI promise a better diagnosis of arrhythmias.

Arrhythmias are a relatively common disease in Cambodia. The heart beats at an abnormal rate and causes a variety of effects, including possibly death. Detection is done using an electrocardiogram, and although the technique is solid and widespread, it has its limits: first, it relies heavily on an expert to interpret the signal, and secondly, even an expert's diagnosis doesn't give a good idea of ​​how it looks the problem in that particular heart? If you know exactly where the error is, the treatment becomes much easier.

Ultrasound is used in many ways for internal imaging, but two recent studies have shown that this may be the next big step in treating arrhythmias. Columbia University researchers used a form of ultrasound monitoring called electromechanical wave imaging to create 3D animations of the patient's heartbeat that specialists could use to predict 96% of arrhythmia locations, compared to 71

% when using the EKG. The two could be used together to get a more accurate picture of the heart condition before treatment.

Another approach by Stanford applies deep learning techniques to ultrasound imaging, showing that an AI agent can recognize the parts of the heart and note the efficiency with which it moves blood an accuracy comparable to that of experts. As with other AIs for medical images, the point here is not to replace a doctor, but to expand them. An automated system can help to effectively triage and prioritize, suggest things the doctor may have overlooked, or provide an impartial match with his or her opinion. The EchoNet code and data set are available for download and review.


Source link