MICCAI 2016 Daily - Thursday

I do not think that protecting someone and hiding her is the way to deal with this. I want her to learn how to deal with this by herself, which is what happened with me: the first time it happens to you, you do not know how to react, and then you have a spontaneous reaction. Even if she’s shy, she has to build that capability MICCAI Daily: Did you ever feel competition and hostility between women? Aïcha: Yes. It funny because Academia is a competitive setting by itself. Everybody is competitive in a lab and people tend to compare themselves to others. So women can be hostile one to another in certain situations, but it is not an issue of gender inequality. Women can be competitive one towards another just like men are. MICCAI Daily: So this is a point which is symmetric. Aïcha: I think everything should be symmetric! This is my point. We shouldn’t analyze women and analyze men, we should be just equal. MICCAI Daily: Please tell me something about your work. Aïcha: OK. This year at MICCAI I am presenting my work on a topology aware deep learning model for multi part object segmentation. For many years now, lots of research has focused on energy-based models and a lot of advances were made on designing and adding priors to these models to better describe objects to segment. However these traditional techniques generally rely on strong data terms which are often sensitive to image appearance. Recently, there has been a turn in the field with many successful applications of deep learning to various medical applications. Despite their success, existing deep learning models do not specifically encode higher order priors such as objects' topology. Our work attempts to combine the best of both approaches: traditional techniques and deep learning models; and we proposed to augment the loss in convolutional networks with topology priors such as smoothness; containment and exclusion. MICCAI Daily: What are your dreams? Aïcha: I am passionate by the medical field and I want to use my skills to help in that sense. I would like to use what I learn in a more applied system, where I can really see doctors using what I am doing. My dream would be to show to doctors, one day, a tool that I have invented or helped creating - and see them using it, helping people with that. That would be a real achievement. MICCAI Daily: You’re a dreamer… Aïcha: In that sense, yes [ laughs ] MICCAI Daily: And a passionate. Aïcha: Yes, I think yes [ laughs again ] CVPR Daily: Thursday Women in Science 16 MICCAI Daily: Thursday