Meet the intern – Stefan van den Berg

Meet the intern - Stefan van den Berg

Every year students from universities in the Netherlands, Belgium and Luxemboug apply for internships at IMAGEM. We bring you to our intern this spring - Stefan van den Berg.

Stefan is an intern at IMAGEM since March 2018. During a five month period, he will research the concept of ‘Machine Learning’. The project entails recognizing Dormers – but automatically via the computer. Read his story below.


Q: Tell us about your academic background

I finished my bachelor’s in social Geography and Planning at the University of Utrecht. During this bachelor I learned about GIS – mainly in my minor National Geo-Information at the VU in Amsterdam – this was very interesting. My internship was done at Kieskompas – an international player in the field of voting statistics. During elections they collect data, which often contains a geographical component. Things you can do with it, for example, in America – are people more conservative in the middle of the country compared to the west coast? All of this could be visualized with GIS.

Currently I am doing my masters, called GIMA (Geographical Information Management & Applications). This is my final year and being an intern at IMAGEM is part of this.

Q: why do you like geography?

The nice thing about geography and mainly GIS is working with spatial data and maps. It offers endless amounts of possibilities and gives you clear results, something I’ve missed during my bachelors with some courses.

Q: You will use ‘Machine Learning’, were you already familiar with this?

Machine Learning is something entirely new to me. I got a better understanding of the concept after talking to someone from IMAGEM – who shed some light on the world of Machine Learning and Deep Learning. Recently, IMAGEM got into Machine learning – quite exciting I can tell you and I all the more reason to dive into this subject.

Q: What entails your research?

The subject of my research was something IMAGEM provided, but after consultation, I adjusted it a little bit. Recognizing Dormers (A roof structure that projects from the roof mass, has its own roof, windows and sometimes walls) automatically is the main subject of my research. Using a Convolutional Neural Network (CNN) deep learning model should provide the wanted results.

Q: How does it work?

Teaching the ‘neural network’ how Dormers look is the first step – this is also known as training data. The next step is applying the trained network on new aerial photos to see what if the network can recognize Dormers automatically or needs more adjustments and/or training data.

Q: How did you end up at imagem?

During GIMA-day, an event of my program, I had a nice conversation with someone from IMAGEM – who is now one of my colleagues at IMAGEM. He gave a cool presentation about 3D-models. My first thoughts were: “wow, what a great and innovative company – it would be amazing if I could do my internship there”. Now, a few months later I am here.

Q: What do you think of IMAGEM as a company?

I think it’s a great company. Some products I was already familiar with, like ERDAS IMAGINE. The great thing about IMAGEM is the small and personal aspect of it. Personal contact is something they value greatly and because of this you get to know people well and it creates a healthy working environment.

Q: tell us more about yourself, your hobbies, likes & dislikes.

I love sports, unfortunately I had to quit playing football because of a knee injury but now I do fitness to stay active. Quite some time is spent on watching football, not just because I like it but for my part-time job I keep track of statistics of football games. Things like the number of faults, goals, who scored and from which position are all stats I keep track off.

Traveling, seeing different cultures and countries is something I really like. Together with my study group I made a trip to Scotland where we did a lot of study related researched – which was both interesting and fun. After I finish my masters, traveling somewhere far-away is high on my bucketlist – most likely South America. The culture, people and different scenery is something fascinating and I would love to see with my own eyes.