Like many great computer science ventures Rithmio was born at the University of Illinois Urbana Champaign. And though that location puts the startup slightly outside our usual reporting focus, the founder’s interest in opening a Chicago office soon, its numerous Chicago-based backers, and its downright fascinating product makes it worth bending the rules.
Rithmio has developed technology that recognizes motion in almost any device, with minimum input from a development team. That means from jumping jacks to running, from iOS to Android, this general-purpose motion recognition software learns human movement on its own.
The software is so free thinking that, “once we[Rithmio] had learned those motions we would be able to recognize when you were doing them again,” said founder Adam Tilton.
To build this software Tilton and his team have been grinding away on complex algorithms that recognize patterns in human movements, and then classify those movements into distinct groups.
“We solved the pattern recognition and classification problem better than had been done before,” said Tilton. “When you apply pattern and classification to motion data what you arrive at is motion recognition.”
Typically motion recognition is done like everything in computer science, by inputting a lot of data and instructions. Making a fitness app? You might get a bunch of users together, have them naturally do a number of physical movements, then record the data generated from those users.
Via such a method, “it might take three months to train the system to recognize a movement,” said Tilton. “And in the real world info might be different than the training.”
“Our device can be trained in as little as six-10 seconds,” said Tilton. So, “you don’t have to worry about solving this very difficult signal processing problem. And you can focus on differentiating your self on other services.”
Rithmio in that sense is a true platform. The software is a jump-start for developers to do more useful things with their time. Instead of spending time programming movements, developers can now work on making use of those recognized movements.
Tilton imagines the software not only recognizing exercise movements, but other slighter gestures. Titlon said the software is so accurate that it could recognize a unique signature performed in the air. Furthermore, it is imaginable that motion recognition software like this might someday be so pliable that everyday users could use the software to recognize their movements, then reapply that data for their own uses.
However, that day is sometime off, and in the meantime Rithmio is building out development and marketing teams. Tilton is spending a lot of time in Chicago looking for additional team members. He imagines the company might have a Chicago-based office where a development team will work on building out SDK’s for iOS and Android.
Those new hires will be paid with funds that Titlon said came mostly from the Midwest, including a lot from Chicago. Rithmio has raised $650,000 in seed funding from Marcin Kleczynski, CEO of Malwarebytes, who led the round, and with participation from other investors that included, Serra Ventures, BonaVentura and Fox Ventures. Chicago-based investors include Illinois Ventures, Hyde Park Venture Partners, and Techra Investments.
“It's really awesome to have been able to have done this with Midwest money,” said Tilton.