Orzata focused on computer vision algorithms
(EDITOR’S NOTE: This is the third in a three-part series spotlighting some of the entrepreneurs participating in the inaugural autoXLR8R in Middle Tennessee.)
Two former Cornell University college classmates, one from New Jersey and the other from St. Louis, have reunited this summer to develop applications that bring visual capabilities to devices like robots and overcome what one of the co-founders describes as his “research frustrations.”
The duo founded Orzata, one of 10 companies that are part of the autoXLR8R housed at the former Saturn Corporation headquarters building in Spring Hill. The company is billed as a “rapid prototyping platform for advanced computer vision algorithms.”
In fact, the start-up is already promoting its first product, a platform technology named Protoboard ™.
How the co-founders reunited is a simple matter of staying in contact after they earned their bachelor’s and master’s degrees in electrical engineering and relocated to different parts of the country. How they chose Middle Tennessee to jump start their company is a matter of convenience and the new autoXLR8R.
Ankur Kumar, the start-up’s Chief Technology Officer, is a doctoral student at Vanderbilt University, focused on image-guided surgery. His co-founder – Jaydev Mahadevan – serves as President and Chief Executive Officer. For several years, the latter worked as a web engineer for start-ups in several cities.
“We want all computer vision of the future to be developed on our platform,” Mahadevan said in a recent interview with teknovation.biz. The duo is positioning the Protoboard ™ system in the software as a service (SaaS) spectrum and is soliciting input on the web page at https://www.orzata.io/.
They are passionate about turning around what Kumar describes as the “development bottleneck common to all sectors,” referring to the cumbersome process of prototyping complex algorithms that frequently takes months to complete. Their goal is to reduce it to a few hours, “ideating solutions faster” to make go or no-go decisions.
So how do robots relate to their interests? It’s all about their long-term focus on image analysis, characterized by Kumar as “making robots (or cars or other programmable devices) see like humans.” To underscore his point, he asked, “Why don’t robots load dishwashers?” The answer was simple – “They can’t see. Robots lack instinct, but can be programmed.”
That’s the intersection with the autoXLR8R, bringing cognitive ability to programmable devices such as cars.
In a recent blog post, autoXLR8R Program Director Mike Nichols wrote about the many opportunities that he sees for Orzata.
“Think about all of the things devices could do if they knew how to see and recognize what was happening around them, and respond accordingly,” he suggested. “The algorithms already exist that instruct a computer on how to do just that, but as consumers we have only seen them in limited use.”
One specific example that Nichols cited was a personal car that could be programmed to recognize the owner as he or she approached and automatically open the door.
Other areas that Kumar and Mahadevan discussed included medical imaging, surveillance, autonomous vehicles (advanced assisted driver systems), image-guided surgery, and even court trials.
“Imagine analyzing blood flow and the muscular aspects of the face of a juror or witness during a trial,” Kumar said.
For those interested in learning more about Orzata, there is the web page, but Mahadevan and Kumar will also present the company at the autoXLR8R’s “Demo Day” on August 7 in Spring Hill.