Computer Vision News - April 2020

3 Summary Stamatia (Matina) Giann rou 1 because we have all these deformations, all these artefacts. For example, we have tools and blood occluding the scene, and we have tool-tissue interaction deforms without following any motion models. Part of my work has focused on estimating tissue deformation. We have proposed a framework which got the Philips IPCAI Best Paper Award at IPCAI 2016.” The next step is putting all the components together and preparing the platform for the operating theatre. What breakthrough would Matina like to see in the future? “What I’d like to see, and it would help a lot with medical imaging, is people not using the diagnosis support platforms like a black box. The surgeons don’t like having an image as input and then a label as output. They want to understand what is going on. I believe interpretability of all these techniques is needed.” Matina concludes by telling us that at MICCAI the numbers of computer- assisted intervention papers could be higher. This may be because it’s a challenging area, she suggests, or they may just need more people to work on it. Her message to those people who are working on it is clear: “It’s nice to come up with blue-sky methods and think outside the box, but we always need to keep translation in mind. We must build something that is ultimately useful for surgeons.” Autonomous Robotic Tissue Scanning. S. Giannarou, M. Visentini-Scarzanella, G.Z Yang, “Probabilistic Tracking of Affine-Invariant Anisotropic Regions”, IEEE Trans. Pattern Anal. Machine Intell., 35 (1), pp. 130-143, 2013 P. Triantafyllou, P. Wisanuvej, S. Giannarou, J. Liu, G.-Z. Yang, “A Framework for Sensorless Tissue Motion Tracking in Robotic Endomicroscopy Scanning”, ICRA2018