Computer Vision News - December 2020

is to i ncorporate more modalities to build a multimodal method to learn the representations. The PULSE dataset includes eye-gaze tracking data and motion data from the sonographer, for example. The team would also like to explore what the model actually learns for this specific representation and interpret how and why it learns such useful representations. 2 Paper Presentation 32 Finally, we are fascinated to know – what is it like to work with such eminent professors as Alison Noble and Andrew Zisserman ? “I learn a lot from them and am very grateful and honored to be advised by them,” Jianbo tells us keenly. “They are both very supportive of my research and give me sufficient freedom. On the one hand, I get the computer vision perspective , and on the other, Best of MICCAI 2020