Computer Vision News - September 2020
355 Emanuele Trucco Best of ECCV 2020 Our collaborators are mostly clinical centers. Our model, at the moment, is a strongly collaborative one. We don't do publicity. The only way people reach us is through our publications, or contacts, or networking, and so on. People who come to us are typically clinicians who are interested in the questions that I was telling you before, who come with their own raw data. We've developed ways of enabling known technical groups, known clinical groups, to use our software to ask questions. I'll give you one example, for instance. We have now a considerable project with a network of diabetes specialist hospitals in India, with the main seat in Chennai. This is an NIHR funded project. It's seven million pounds for five years, and we're basically developing tools for precision medicine for diabetes. This includes genomics, genetics and therapies, and retina imaging analysis. It's exciting because, in addition to developing new tools, we’re working with people with a different perspective or with a different scale of the problem. You can imagine Scotland versus rural India, where the need for telemedicine is strong, and we have access to half a million records in retinal images. Also, this is projected on a scale, which is fantastic. In thinking of large data sets, we also work in the UK with the only guy on the block, UK Biobank: as VAMPIRE, we have been members of the UK Biobank Eye and Vision Consortium, basically, nearly from inception. It is a very, very interesting time. The tools we develop, as you can imagine, revolve around image analysis, artificial intelligence, deep learning. We have very exciting things happening in two further fields at this moment beyond the eye. One is dermatology in Scotland, and the other is neurology. These are hot off the press within the last six months to one year. For dermatology, we've developed a pilot in Dundee, which is now going into a larger proposal, which might turn into a Scottish-wide instrument for triaging lesions. And the other, for neurology, is a collaboration with the University of Glasgow, who has a fantastic growing dataset of smartphone videos collected by the public of epileptic seizures. Andthis is a fantastic dataset for developing A.I. techniques, for helping people discriminate between these types of seizures. So, these are exciting times. You also organized a workshop called OMIA at MICCAI, which is centered around the eye and the retina. What
Made with FlippingBook
RkJQdWJsaXNoZXIy NTc3NzU=