Career story: Imaging Software Engineer Eero

9 Mar 2021

Developing own imaging algorithms helped to finalize Master´s thesis and later to “see” own work result in image quality

It was 2017 spring when I first heard from Visidon. I was a computer science student at University of Oulu, trying to finish my master’s degree and looking for a summer job. My professor asked if I was still looking for a job. Well I was, I had not gotten an answer from most of the places I sent an application. He knew Visidon because the 3 founders had previously been working at the University of Oulu (and even achieved PhD’s), as I learned later.

Back then Visidon was a small company (about 7 workers) that had just started to grow. The interview was pretty chill and I didn’t think much of it. To my glad surprise, they called me back next week and offered a job! I gladly took it and after the summer I was hired as a full-time worker.

One of my first tasks was to familiarize myself into multi-frame imaging, de-noising and HDR. I implemented some prototypes with Matlab and tried to speed-up with C/C++. We were mostly interested about the quality and how fast it would work on Android smartphones. The Fast Fourier Transform the algorithm was using was a bit too slow on smartphones back then, but at least I learned a lot about C coding from the task.

At the university I had been mostly coding with Python and knew enough about C/C++ to run simple OpenCL or CUDA codes with some OpenCV. The algorithms themselves are not rocket science. My university’s image processing courses prepared me well for my tasks: blurring, sharpening, rotating or gradients are simple but often used processes in image manipulations. The problems usually comes from speed or when applying the same algorithm to various cases e.g. night versus daytime pictures.

I did say I was finishing my master’s degree when I took the job. But for some reason my first thesis subject received from the University did not end up working, so I decided to check if one of my algorithms developed here at Visidon could be used for finalize my thesis. Luckily I had been working with some novel chromatic correction algorithm which I was able to use in my thesis. And in this way I received my MSc.

“Blurring, sharpening, rotating or gradients are simple but often used processes in image manipulations. The problems usually come from speed or when applying the same algorithm to various cases e.g. night versus daytime pictures.”

The hierarchy in Visidon is horizontal, meaning I’ve always been able to ask about anything from anyone if I have had a problem. This was especially helpful during my first summer here. Some of the stuff here is not widely used in other fields (e.g. YUV image format), but there was always someone who could explain it to me. Even now if I get stuck in my coding I might just get up and ask help from a coworker. It’s surprising how many talented or knowledgeable people we have found and I’m sure that at least one of them knows the answer. Now that we are a little bit bigger company, we try to keep it that way and not get stuck in our own little groups. Although I do feel like an university researcher sometimes, developing my own algorithms (minus the scientific paper writing). Overall I think we are still an usual tech company. We even have a ping pong table like all the cool startup companies.

Personally I think that Visidon is the perfect place for me. With image processing you can literally “see” your work and there’s a million things you can do with an image. My coworkers are nice too and during lunch we joke a lot at the lunch table (which sadly has quieted down during the pandemic).

Feel like joining the VD team? Apply now!

See our open positions or send us an open application: