Computer Vision News - April 2023

15 Seyran Khademi Do you have a preference between one world and the other? Well, my comfort zone is actually computer science because I've been there almost forever now. So that's a comfort zone. It's not a preference per se, but I think architecture is also fascinating, especially the people working in this domain: they are kind of mediators between us, nature, and our built environment. They have to develop lots of skills in lots of directions, from art to engineering, to social science. And that makes a very interesting profile. For me, being in this context, it's a very valuable experience to learn from different domains. I think, in general, we are kind of moving away from a single discipline. Our problem as humanity has become so complex that it really needs to be solved at the intersection of many disciplines. When you become very famous, we will see your name on a building. Or maybe there will be a computer program with your name. We don’t know yet. [ laughs ] Very interesting! Projecting to the future, I would not know, actually. But I think both are fascinating. Tell me Seyran, if you were not working in this field, what would you do? What would your career have been? Very interesting question! You're taking me to the parallel world of me being something else [ laughs ] This is very unimaginable, but I think I would still be teaching. I think the teaching part, apart from the research, is very rewarding forme. I would probably still work in some kind of academic education institute, and Iwouldbe teachingsomething on science still. I cannot imagine working outside of science. For me, it's again a safe zone, a haven that can help me understand Seyran, what do you do at Delft? I came to Delft in 2011 to do my PhD in signal processing. Signal processing is between mathematics and data science. It's the science behind our digital life, basically. I did my PhD for four to five years at the Department of Signal Processing in the signal processing lab. Then I moved to computer vision as a postdoc. In between, I did a bit of speech processing as well. So it's been a journey for me to come from low- level signals to really complex signals like images, videos, and any kind of visual data. Right now, I amworking on AI for visual data processing at the Faculty of Architecture and the Built Environment. As architecture is by nature a visual discipline, designers and architects communicate their design ideas very much with visual data. And with the whole era of big data, we need intelligent models to be able to process all this kind of data and tomake sense of them. What I'm doing right now is developing and working on AI methodologies for architects to better explore their visual domain. Our readers know already one lady who works at the intersection between architecture and computer vision, Iro Armeni. Maybe you have heard about her? Yeah, I know her! I'm actually kind of collaborating with her right now. [ laughs ] It's a small world after all! [ laughs ] Yes, indeed. In this intersection, it is. I think the difference is that she actually came from the world of architecture to computer science. And for me, it's the other way around. I started as a computer scientist and moved to the domain of architecture.