Sinisa Todorovic admits that on the surface his two careers — ballet and artificial intelligence — seem unrelated. “I sometimes wonder why I trained for hours!” he jokes of his ballet career that began when he was just 10 years old.
Selected to dance in the premier troupe in Serbia, he travelled the world as part of an entourage for the Serbian president, Josip Broz Tito, to represent their country.
“We used to be welcomed by thousands and thousands of people cheering us. It was pretty spectacular,” he says, recalling that he had the privilege of meeting many of the royal families in Europe and traveled twice to North Korea.
He continued dancing until age 23 when a series of civil wars in the 1990s and the resulting lack of a performing arts culture in Belgrade made it impossible to have a ballet career. So instead he pursued a degree electrical engineering at the University of Belgrade in Serbia.
Although life was very difficult then — he lived with his parents who at the time were earning the equivalent of $1 a month and were slowly selling off all their belongings to survive — he also remembers it as the best time of his life.
“Usually in war times people under oppression tend to be more forgiving, more caring, and so there was just an amazing spirit at the time,” he says describing how during bombing raids the women would dress up to go to the bomb shelters as if it was a party. Power outages could last for months so people went back to simple entertainment like reading and talking to each other.
Still, when he had the chance to get out, he did. As soon as the borders opened he got a job as a software engineer with Siemens and worked in Vienna and Moscow before coming to the US for graduate school and receiving his PhD at the University of Florida in electrical and computer engineering. He specialized in computer vision and received a prestigious post-doctoral position at the Beckman Institute at University of Illinois at Champaign-Urbana before coming to OSU.
“Artificial intelligence really combined all my loves — philosophy, psychology, and math — everything that I really wanted to do,” he says.
He explains that in order to detect specific objects in an image, such as a face, engineers are challenged with really fundamental questions.
“I was so happy when I took my first class in computer vision, and the professor asked us, ‘What is a face?’ You don't hear these kinds of questions in other classes,” he says.
Face detection is important for applications such as automatic focus on digital cameras. Some of his more recent research aims to track moving objects in video, such as football or basketball players to potentially automate play-by-play transcription of sports videos.
He is also working on a project that incorporates yet another passion — movies. His dream would be to work with movie companies to extract objects from old video and combine it with new footage, so that Marilyn Monroe could be in a scene with Brad Pitt, for example.
Although being a ballet soloist was one of his proudest achievements he ranks just as high the accomplishment of nurturing his graduate students from fledgling researchers to accomplished professionals.
And although he no longer dances, the hours of practice he endured were not wasted because it taught him the dedication and discipline that is required to be good at something.
“What is important is to find passion in whatever you do — whether it is dancing or computer science. Any career offers its challenges. So, whatever you do put all of your effort into it,” he says.