Computer Vision News - June 2020

3 Summary WEISS at UCL (part 2) 2 emerging scientific discipline with the objective of improving the quality of interventional healthcare and its value through capturing, organization, analysis, and modeling of data. She is keen to mention Lena Maier-Hein , who she declares is leading the field. “It’s challenging at the moment because it’s not easy to get data from hospitals, but I think it’s going to be a very promising field that will see many advances in the coming years,” Ester says optimistically. “There are many people focusing on this. They want to integrate all the knowledge that the hospital has together with imaging to help with decision support and other clinical tasks. I’m particularly interested on the image side. How can we help gather more data from imaging?” In the case of ultrasound , it’s even more challenging because ultrasound images or videos are not stored by default. You need to acquire more data and use this data quickly to impact the clinical outcome. Another challenge is getting labelled data from ultrasound images, which have information embedded that can be used to train new methods. Ester is hopeful: “I’m working on it!” Thinking about her own future work, Ester would like to focus on data- driven interventional healthcare . “I’m interested in creating methods that are able to learn frommultiple sources of data,” she says excitedly. “Mainly imaging, because it ’s my expertise, but also from multi-source tracking and things like that.” Also, she would like to create context-aware software that is able to understand what is happening during an intervention or procedure. This kind of software would allow for decision support focused on data from the procedure itself, which she hopes will allow them to access more data, better integration and better generalization of core methods. “There’s something called federated learning that I’m very interested in too,” Ester adds. “It ’s about gathering data and then putting the trained data in the cloud. I have a particular interest in methods to join different sources like that.”

RkJQdWJsaXNoZXIy NTc3NzU=