CogniVue has only been in business for four years, but the company’s roots go back more than a decade.
A depiction of Cognivue's APEX parallel processing system.
By Jacob Serebrin
Its history can be traced back to the founding in 1999 of Atsana Semiconductor, which was purchased six years later by South Korean semiconductor manufacturer MtekVision. It named the organization, then an R&D facility, CogniVue. MtekVision spun off CogniVue in 2010 with the goal of turning an R&D cost centre – a facility that doesn’t produce any direct revenue – into a profitable corporation.
Despite that history, the company’s CEO, Simon Morris, says CogniVue has a startup attitude.
“We’re a bootstrapped company, like a startup, but we’ve got an experienced team,” he says.
CogniVue is already seeing success with its core offering, an embedded vision processing chip. It’s a tiny computer system that detects objects, even if they’re moving, designed to go inside small cameras. The system, which CogniVue calls an image cognition processor, is being used in “smart” rear-view cameras for cars.
When the product launched, it was the first of its kind.
“We had a disruptive advantage,” says Mr. Morris.
One of the key factors, he says, is that the chip runs cooler than its predecessors. It’s an important factor in the small, sealed cameras used in the automotive sector because the heat would build up and distort the image, Mr. Morris says.
But developing a system that could process the large amounts of data generated by raw video without overheating wasn’t easy.
“The biggest issue is how you organize the data,” he says. “It was all about thinking smart about minimizing data movement.”
The chip is also smaller, which reduces costs for CogniVue and keeps prices down.
Mr. Morris sees even more opportunities for his company beyond the automotive sector.
In December, the company signed a deal to license its technology to Tokyo-based Digital Media Professionals, which develops graphic processing units for Nintendo and other companies.
There are cameras everywhere, but they don’t have any intelligence. Simon Morris, CEO of Cognivue
Mr. Morris says the chip also has growing applications in the mobile and wearables – things like Google Glass – markets, where the chip’s efficiencies translate to better battery life.
“This is stuff that’s not really mainstream yet but will be in a few years.”
Mr. Morris says he also sees new revenue opportunities as the chip becomes more widespread.
He’s looking to build a “parallel revenue stream where we license our software to our customers’ customers.”
He says this would speed up the process of developing applications for the camera.
“We see that part of the business as potentially more lucrative,” he says. But before the firm can take advantage of that channel, Mr. Morris says his team will have to build up the market. So for the next three years, CogniVue’s primary focus is licensing the processor technology to as many chip manufacturers as possible.
That’s not only the key to the company’s future plans; it’s also its key source of revenue.
“There’s lot of execution challenges, but that’s the vision,” he says.
Head count: 25 full-time, 8 part-time
Funding: Undisclosed venture funding “in the millions” on top of seed funding from MtekVision Also funded through grants, loans and tax credits along with revenue from sales and licensing.
Clients: Original equipment manufacturers in the automotive sector
Product: A computer chip for processing and understanding visual data