Computer Vision News - December 2020

2 Medical Imaging Technology Talks 34 David Menashe is an algorithm team leader at RSIP Vision. He writes about our market-leading work using AI to enhance medical ultrasound applications. Ultrasound is an easy-to-use, portable , and inexpensive technology, with many different applications, making it one of the most promising technologies in medical imaging. An excellent current example of the versatility and importance of the technology is the recent application of lung Ultrasound to COVID-19 diagnosis, and specifically to rapidly and effectively identifying severe cases. However, ultrasound images canbe noisy and are not always easy to interpret. It takes a lot of skill by the technician to produce a good image and even more skill to interpret images or videos and derive the correct diagnosis. In heart echo studies, for example, there may be 60-70 video clips of the heart and nearby blood vessels to analyze, with various calculations and measurements to make to obtain a full picture of what is going on. Cardiologists use these clips to diagnose a multitude of different diseases and conditions, which can be monitored and followed up at regular intervals. It is a time-consuming process when performed manually , and often clips can be irrelevant or of poor quality. This is where artificial intelligence steps up. AI can help experts to sift through a huge amount of data, identify critical images or clips, and eases the workload for medical professionals involved in diagnosis. At RSIP Vision recently, we have used AI to enhance the analysis of heart echo studies, including analyzing the video clips to assist the physician to select only those which are good quality; sorting the clips according to what view and cross- section of the heart they show; detecting features on the clips, such as valves and chambers, and automatically taking measurements, such as one of the key measurements in cardiology, the ejection fraction, which measures how efficiently by David Menashe Best of MICCAI 2020