Computer Vision News - February 2020

2 Summary Medic l Imaging 8 Leo Joskowicz is a Professor at the School of Computer Science and Engineering at the Hebrew University of Jerusalem, Israel. He is also the Director of the Computer Aided Surgery and Medical Image Processing lab (CASMIP). On top of all that, he is the President of the MICCAI Society and is currently organising MICCAI 2020, which will be held in Lima, Peru in October of this year. He has taken some time out of his busy schedule to talk to us about his work and the latest trends and innovation in AI for medical imaging. This is an exciting time for people involved in medical image computing and computer-aided surgery. Over the last few years, the need for workf low optimization in radiology has grown and so have the opportunities. The use of CT and MRI scans has increased, growing 10 to 15 per cent faster than the number of radiologists who are needed to interpret the images. The untapped information in those scans could be utilised for many important functions, including diagnosis, treatment planning, treatment delivery, and treatment evaluation. Now, with the advent of deep learning techniques in AI, we are able to do things that we were unable to do without considerable effort, or at all, before. Leo tel ls us that the “low-hanging fruit” for AI in radiology is in detection and classi f ication. Deep learning technologies have al lowed the bui lding of detection and classi f ication systems in a way that could not have been done f ive or ten years ago. There are a number of companies whose business model is based on processing an X-ray, CT or MRI scan and being able to tel l a radiologist or cl inician i f there’s something there. That something could be a haemorrhage or a tumour, for example. It ’s a warning system to alert the cl inician to take a closer look and could reap benef its for patient triage. The cl inician might not be looking for a brain haemorrhage, but i f they already have a CT or MRI of the brain, why not run it through an algorithm and see what comes up? Listening to Leo talk, you could forget that he began working in the f ield of medical imaging and computer-aided surgery almost